Pinned Tweet
My pinned tweet explains the background to Healey resignation.
We got real numbers in 2020 for first time in ~20 years.
Since then everything reverted to fake.
Thanks to Trolley & Wally, they let HMT and MoD shovel cash into things they KNEW were broken, like AJAX, and scuppered fundnes, robotics, AI, cloud).
The nuclear weapons supply chain is a multi-billion multi-decade fiasco and reinforces fake budgets.
MoD & CO used UKR war to prop up the lies and disasters, not to fix them. Trolley, Sunak and Starmer all went along with: classify, punt, spin fake budgets.
All these chickens are coming home to roost.
The DIP is a disaster because it combines: a/ continuing to fund old things which shd be scrapped but senior people's careers rest on lying about, b/ failure to fund the future, c/ more dodgy accounts, d/ classified nuke shitshow, which forces lies/cannibalisation of conventional, and e/ continued failure to change procurement and long term budgets despite covid and UKR.
Ive said this 100 times for years.
Also relevant to resignation --
In 2020, I got the PM & CHX to agree to take the next gen fighter AWAY from HMT and MoD. (FOI the doc screenshotted below!)
They were using a bogus 'above STRAP 3 study' to claim that AI would be irrelevant to future fighters in 2045. This 'top secret' study was written *before* DeepMind was created!!
We agreed to take away the next gen fighter/drones from MoD/HMT and put them into a new civilian project with huge commercial development of drones/robotics/AI, with procurement and budgets done totally outside normal system (like vaccines, ARPA etc).
After I left this was u-turned along with shredding all the true budget numbers.
The MoD and HMT proceeded with the usual procurement with BaE.
This has now blown up *as it was obvious it would SIX YEARS ago'.
HMT is refusing to fund it because the MoD is a corrupt shambles.
And MoD is knackered by the HMT's insane budget processes and insane procurement.
HMT and MoD have reasonable complaints about each other. 'Everyone's right and everyone's unhappy'.
The only way out is the Vote Leave plan and No10 smashing heads together with the full authority of the PM, as in the secret process of 2020. It cannot be left to 'normal processes'. Senior people must be forced to tell the truth 'or you're fired'. We did this. It works!
The MPs refuse to engage. The MPs on defence committee say they cannot ask officials re the 2020 process because 'they'll just refuse to answer'.
But our enemies know everything everything I say is true!
It is impossible for dud Starmer to solve these problems.
It is good timing that Burnham will soon take over and could draw a line under decades of Tory failure and set out a new path...
But there will be HUGE pressure from the old system, in HMT and MoD, to patch everythign up with more lies, spin, fake accounts, and massive classification...
Maybe some MPs could finally pay attention to these issues and resolve to stop constant lies, please!
Our security, prosperity and critical technologies are at stake and the old parties and old Whitehall are pathological!
NB. In 2020 we also bought LEO satellites. SW1 laughed. Starlink has been critical in UKR. Whitehall sold off the satellites!
NB. In March 2022 when the UKR disaster started I wrote: 'Prediction: 1) lessons from UKR will overwhelmingly support the arguments of those who in 2020 argued for radical MoD changes (including taking money from old tank projects that everybody privately admitted were a multi-billion pound disaster) and 2) the correct criticism of the review and connected documents will be seen as a) they did not go nearly far enough, b) the collapse of No10 follow through on defence reform in 2021 was — like the collapse of 2020 plans for planning reform, tax cuts, deregulation, Project Speed, intense focus on R&D and skills etc — a disaster for the country (and a political disaster for the Tory Party).'
NB. Special Forces who have observed the drone wars in UKR have told the senior MoD repeatedly that the mega expensive things they want to plough cash into in the DIP will get VAPOURISED in a real war by cheap drones! The SF are told to shut up and not give HMT a chance to cut the budgets!
I know it's psychologically extremely hard for SW1 but Vote Leave worked with the deep state on this in 2020 and we were right on all the big things.
The disastrous old processes have now blown up with unprecedented resignation over these exact issues.
6 years have been wasted. Tens of billions have been wasted. Lives lost. Talent resigned.. Opportunities to lead squandered...
All over Whitehall are officials involved in 2020 who know what I say is true, the MPs and hacks should finally get the details into the public...
Maybe SW1 could update its mental models?!
Ps. if you're a hack still getting Ciaran 'AI is fake, PRC infiltration is overrated' Martin to comment on tech, you're as dud as Starmer
We got real numbers in 2020 for first time in ~20 years.
Since then everything reverted to fake.
Thanks to Trolley & Wally, they let HMT and MoD shovel cash into things they KNEW were broken, like AJAX, and scuppered fundnes, robotics, AI, cloud).
The nuclear weapons supply chain is a multi-billion multi-decade fiasco and reinforces fake budgets.
MoD & CO used UKR war to prop up the lies and disasters, not to fix them. Trolley, Sunak and Starmer all went along with: classify, punt, spin fake budgets.
All these chickens are coming home to roost.
The DIP is a disaster because it combines: a/ continuing to fund old things which shd be scrapped but senior people's careers rest on lying about, b/ failure to fund the future, c/ more dodgy accounts, d/ classified nuke shitshow, which forces lies/cannibalisation of conventional, and e/ continued failure to change procurement and long term budgets despite covid and UKR.
Ive said this 100 times for years.
Also relevant to resignation --
In 2020, I got the PM & CHX to agree to take the next gen fighter AWAY from HMT and MoD. (FOI the doc screenshotted below!)
They were using a bogus 'above STRAP 3 study' to claim that AI would be irrelevant to future fighters in 2045. This 'top secret' study was written *before* DeepMind was created!!
We agreed to take away the next gen fighter/drones from MoD/HMT and put them into a new civilian project with huge commercial development of drones/robotics/AI, with procurement and budgets done totally outside normal system (like vaccines, ARPA etc).
After I left this was u-turned along with shredding all the true budget numbers.
The MoD and HMT proceeded with the usual procurement with BaE.
This has now blown up *as it was obvious it would SIX YEARS ago'.
HMT is refusing to fund it because the MoD is a corrupt shambles.
And MoD is knackered by the HMT's insane budget processes and insane procurement.
HMT and MoD have reasonable complaints about each other. 'Everyone's right and everyone's unhappy'.
The only way out is the Vote Leave plan and No10 smashing heads together with the full authority of the PM, as in the secret process of 2020. It cannot be left to 'normal processes'. Senior people must be forced to tell the truth 'or you're fired'. We did this. It works!
The MPs refuse to engage. The MPs on defence committee say they cannot ask officials re the 2020 process because 'they'll just refuse to answer'.
But our enemies know everything everything I say is true!
It is impossible for dud Starmer to solve these problems.
It is good timing that Burnham will soon take over and could draw a line under decades of Tory failure and set out a new path...
But there will be HUGE pressure from the old system, in HMT and MoD, to patch everythign up with more lies, spin, fake accounts, and massive classification...
Maybe some MPs could finally pay attention to these issues and resolve to stop constant lies, please!
Our security, prosperity and critical technologies are at stake and the old parties and old Whitehall are pathological!
NB. In 2020 we also bought LEO satellites. SW1 laughed. Starlink has been critical in UKR. Whitehall sold off the satellites!
NB. In March 2022 when the UKR disaster started I wrote: 'Prediction: 1) lessons from UKR will overwhelmingly support the arguments of those who in 2020 argued for radical MoD changes (including taking money from old tank projects that everybody privately admitted were a multi-billion pound disaster) and 2) the correct criticism of the review and connected documents will be seen as a) they did not go nearly far enough, b) the collapse of No10 follow through on defence reform in 2021 was — like the collapse of 2020 plans for planning reform, tax cuts, deregulation, Project Speed, intense focus on R&D and skills etc — a disaster for the country (and a political disaster for the Tory Party).'
NB. Special Forces who have observed the drone wars in UKR have told the senior MoD repeatedly that the mega expensive things they want to plough cash into in the DIP will get VAPOURISED in a real war by cheap drones! The SF are told to shut up and not give HMT a chance to cut the budgets!
I know it's psychologically extremely hard for SW1 but Vote Leave worked with the deep state on this in 2020 and we were right on all the big things.
The disastrous old processes have now blown up with unprecedented resignation over these exact issues.
6 years have been wasted. Tens of billions have been wasted. Lives lost. Talent resigned.. Opportunities to lead squandered...
All over Whitehall are officials involved in 2020 who know what I say is true, the MPs and hacks should finally get the details into the public...
Maybe SW1 could update its mental models?!
Ps. if you're a hack still getting Ciaran 'AI is fake, PRC infiltration is overrated' Martin to comment on tech, you're as dud as Starmer
I've explained this ~20 times since 2021.
In 2020 in the secret review of the MoD -- conducted in No10 with officials & without the involvement of the SoS because he was seen as a security risk to the process -- the MoD ADMITTED that AJAX was a 'total shitshow and dangerous to troofen them'. They also agreed to close AJAX.
AJAX was one of DOZENS of things the MoD admitted in this process were disasters and shd be closed down.
We agreed to:
1/ Close the disasters.
2/ Change procurement and MoD budget processes fundamentally.
3/ Publicly explain the lies of the budget process of the previous 20 years and publish for the first time in 20 years budget numbers agreed by MoD, Cabinet Office, and HMT were TRUE.
4/ Shift new money to new programs done with new procurement. E.g drones - which in 2020 Wally thought were fake and he just wanted, like most of the senior generals, more bigger tanks. (When I pushed through buying LEO satellites with MI6 help, much of senior MoD opposed that too. Until Starlink was used in UKR, conventional wisdom in SW1 dismissed it -- look at all the laughter in 2020 when I forced this through.)
5/ Start honesty on the massive classified scandalous shitshow of the nuclear enterprise, tens and tens of billions hidden.
6/ Stop the MoD doing the new fighter using the old procurement system, partly based on 'above STRAP 3' study proving AI /drones would be irrelevant for advanced manned ('crewed'!) fighters until at least 2045. This 'above STRAP 3' MoD study turned out - when I finally got a copy - to have been conducted in ... 2010. I.e before DeepMind was sold to Google. Ludicrous. Normal. I got PM and CHX agreement to do the new plane project OUTSIDE the MoD with a totally new process.
This was all finally agreed a fortnight or so before I left No10.
After I left No10, documents were shredded in <24 hours.
The true budget numbers were destroyed.
Everything shifted back to the fake budget numbers.
AJAX and dozens of other disasters were funded MORE.
The procurement system continued and continued through UKR war. (Today special forces keep pleading with MoD to do drones differently, MoD continues the pathological system and orders SF not to inform ministers of the truth on drones in UKR.)
The nuclear shitshow was classified and punted again.
MoD was given the new plane project which, surprise surprise, is the normal process with normal results.
And everyone agreed that the 2020 secret process was never to be discussed or even acknowledged to exist (even though it was officially briefed it was happening and you can see it on the internet).
Trolley and Wally went along because they're both pure products of the pathological old system: lies and total failure are just another day at the office.
They immediately reverted to the Osborne/Heywood/Mcpherson double counting budget lies which are normal in SW1 and wd get you jailed outside it.
I've said this repeatedly.
The media and MPs will never ever cover it.
Just like the official Inquiry on covid in 2020, so the official truth on MoD in 2020 is lies all the way down.
Instead we will continue to have these fake 'revelations' and 'experts' will continue to say 'I'm shocked, shocked about the latest MoD shitshow'.
One MP on the Defence committee called me earlier this year for advice. I said hold hearings and call Simon Case, Lovegrove and the ~20 who did the secret process to give evidence under oath.
MP: 'That's impossible, they always just refuse to talk to MPs on the grounds of national security.'
Yes, quite. So MPs should either enforce change or quit and do something else with their lives.
What's the point of being an MP in such a corrupt shitshow?
Why bother with your Potemkin 'committees'?
Ps. If you are a soldier injured in AJAX, then bear in mind that *the MoD KNEW you would be injured and did it anyway rather than admit their failures publicly.*
Everyone involved will be given gongs and many a job in BAE.
Gongs all round, the system working as intended...
This is about the 20th time I've explained all this. Watch how the legacy media just keeps wandering around saying 'Im shocked to see this' and does zero probing of the 2020 secret review.
Fixing the MoD is an example of *a regime-complete problem*.
66
425
1,493
494,398
Dominic Cummings retweeted
Correct.
Using Starlink, a civilian commercial service, for offensive drone strikes to kill people has never been allowed, nor should it be.
Starshield, a different system, is controlled by the US government. What they decide to do is up to them.
Ukraine military violated Starlink terms of service and used it for offensive drone strikes to kill people. That loophole was closed last week.
Starlink remains critical to the defense of Ukraine, as it is the only system that cannot be jammed or blown up, at least for now. Were Starlink to be turned off, the Ukrainian front line would collapse immediately. This is NOT a threat, it is simply a statement of fact.
Starlink was also crucial to preventing Russia from overrunning Ukraine in the early days of the war, as all other means of communication were destroyed or jammed by Russia.
SpaceX has done more to defend Ukraine than any private company on Earth BY FAR.
That said, I have nothing but UTTER CONTEMPT for the virtue-signaling evil warmongers who push for endless slaughter and offer no realistic solution to the conflict. May you burn in hell.
End of story.
Using Starlink, a civilian commercial service, for offensive drone strikes to kill people has never been allowed, nor should it be.
Starshield, a different system, is controlled by the US government. What they decide to do is up to them.
Ukraine military violated Starlink terms of service and used it for offensive drone strikes to kill people. That loophole was closed last week.
Starlink remains critical to the defense of Ukraine, as it is the only system that cannot be jammed or blown up, at least for now. Were Starlink to be turned off, the Ukrainian front line would collapse immediately. This is NOT a threat, it is simply a statement of fact.
Starlink was also crucial to preventing Russia from overrunning Ukraine in the early days of the war, as all other means of communication were destroyed or jammed by Russia.
SpaceX has done more to defend Ukraine than any private company on Earth BY FAR.
That said, I have nothing but UTTER CONTEMPT for the virtue-signaling evil warmongers who push for endless slaughter and offer no realistic solution to the conflict. May you burn in hell.
End of story.
STARLINK WORKS THE SAME WAY ITS ALWAYS WORKED. ITS NOT A WEAPONS SYSTEM, NO MATTER HOW MUCH UKRAINE WANTS IT TO BE.
9,340
25,066
149,767
7,958,747
Dominic Cummings retweeted
The decline of the hedgehog is mainly caused by the huge recent expansion of the badger population. In both time and place there is a strong negative correlation between badger numbers and hedgehog numbers.
The scientific evidence that an increase in Mr Brock may well be the chief cause of Mrs Tiggywinkle’s demise is overwhelming.
There’s the direct evidence – the hollowed-out prickle-bearing skins left behind after a badger eats a hedgehog. Only badgers can open a hedgehog after it rolls up in defence with prickles on the outside. Foxes and dogs cannot. Badger predation is often found to be the main cause of hedgehog death in studies in this country.
There’s the timing evidence. The decline in hedgehogs coincided with the rise in badgers since the latter’s protection began in the 1970s and was strengthened in the 1990s. Anecdotally, there are clear-cut cases to show this is likely to be cause and effect, not coincidence. Kew Gardens had a thriving hedgehog population until the mid 1980s when badgers first arrived there. Today, says the Royal Botanic Gardens’ website, “Kew is teeming with badgers”. There are now more than 20 badger setts and hedgehogs are hardly ever seen, and if they are don’t last long. Habitat change and development cannot be the cause in Kew, because there has not been any.
There’s the spatial evidence. Throughout England hedgehogs (great worm eaters) have vanished from worm-rich pastureland, where badgers thrive, and cling on only in suburban gardens and amenity grassland, where badgers so far have not penetrated much. Even here, as one study put it: “the probability of hedgehog occurrence in suburban habitats declined towards zero in areas of high badger density.” As a general rule, wherever badgers are present at a density higher than about one main sett every four square kilometres, hedgehogs are absent. In most of rural England, sett density is two or three times higher.
There’s the behavioural evidence. Back in the 1990s, ecologists discovered that the smell of badgers alone is enough to deter hedgehogs from entering an area. And hedgehogs released into areas of excellent habitat high-tailed it out of there if they smelt badger. They do not wait to be eaten.
There’s the international evidence. Sweden has also blamed its decline in hedgehogs on rising badger numbers. In the Netherlands, scientists reported this yearthat hedgehogs are now confined mostly to urban areas, despite roads and development, because of the spread of badgers in rural areas.
There’s the experimental evidence, which is the clincher. The randomized badger control trials, carried out between 1998 and 2006 in 30 different parts of the country, provided an opportunity to see what happened to hedgehog numbers if badger numbers were reduced.
In pasture land, where hedgehogs are mostly extinct, there was of course little effect. But in “amenity grassland” – parks, lawns, gardens, playing fields etc – there was a dramatic explosion in hedgehog numbers in those areas where badgers were culled: hedgehog density almost trebledfrom an average of 0.9 per hectare to 2.4 per hectare in just five years. By contrast in the control areas where no badgers were culled, the hedgehog density remained unchanged at 0.3 per hectare over the same period.
So we know one way to increase hedgehog numbers: reduce badger numbers. No other approach has been demonstrated to have such an effect. Given that badgers are highly effective predators of bumble bees (those who try to study bumble bees complain that most of the nests they find get dug out and eaten by badgers) as well as of ground nesting birds, and that they carry bovine tuberculosis, there is little doubt that lower badger densities would be good for the ecology of the British countryside as well as for farming.
There is nothing natural about the density at which badgers and foxes live in Britain today.
The scientific evidence that an increase in Mr Brock may well be the chief cause of Mrs Tiggywinkle’s demise is overwhelming.
There’s the direct evidence – the hollowed-out prickle-bearing skins left behind after a badger eats a hedgehog. Only badgers can open a hedgehog after it rolls up in defence with prickles on the outside. Foxes and dogs cannot. Badger predation is often found to be the main cause of hedgehog death in studies in this country.
There’s the timing evidence. The decline in hedgehogs coincided with the rise in badgers since the latter’s protection began in the 1970s and was strengthened in the 1990s. Anecdotally, there are clear-cut cases to show this is likely to be cause and effect, not coincidence. Kew Gardens had a thriving hedgehog population until the mid 1980s when badgers first arrived there. Today, says the Royal Botanic Gardens’ website, “Kew is teeming with badgers”. There are now more than 20 badger setts and hedgehogs are hardly ever seen, and if they are don’t last long. Habitat change and development cannot be the cause in Kew, because there has not been any.
There’s the spatial evidence. Throughout England hedgehogs (great worm eaters) have vanished from worm-rich pastureland, where badgers thrive, and cling on only in suburban gardens and amenity grassland, where badgers so far have not penetrated much. Even here, as one study put it: “the probability of hedgehog occurrence in suburban habitats declined towards zero in areas of high badger density.” As a general rule, wherever badgers are present at a density higher than about one main sett every four square kilometres, hedgehogs are absent. In most of rural England, sett density is two or three times higher.
There’s the behavioural evidence. Back in the 1990s, ecologists discovered that the smell of badgers alone is enough to deter hedgehogs from entering an area. And hedgehogs released into areas of excellent habitat high-tailed it out of there if they smelt badger. They do not wait to be eaten.
There’s the international evidence. Sweden has also blamed its decline in hedgehogs on rising badger numbers. In the Netherlands, scientists reported this yearthat hedgehogs are now confined mostly to urban areas, despite roads and development, because of the spread of badgers in rural areas.
There’s the experimental evidence, which is the clincher. The randomized badger control trials, carried out between 1998 and 2006 in 30 different parts of the country, provided an opportunity to see what happened to hedgehog numbers if badger numbers were reduced.
In pasture land, where hedgehogs are mostly extinct, there was of course little effect. But in “amenity grassland” – parks, lawns, gardens, playing fields etc – there was a dramatic explosion in hedgehog numbers in those areas where badgers were culled: hedgehog density almost trebledfrom an average of 0.9 per hectare to 2.4 per hectare in just five years. By contrast in the control areas where no badgers were culled, the hedgehog density remained unchanged at 0.3 per hectare over the same period.
So we know one way to increase hedgehog numbers: reduce badger numbers. No other approach has been demonstrated to have such an effect. Given that badgers are highly effective predators of bumble bees (those who try to study bumble bees complain that most of the nests they find get dug out and eaten by badgers) as well as of ground nesting birds, and that they carry bovine tuberculosis, there is little doubt that lower badger densities would be good for the ecology of the British countryside as well as for farming.
There is nothing natural about the density at which badgers and foxes live in Britain today.
Go for a walk on almost any farm in Britain and ask yourself where a hedgehog might find food, water or shelter there. The rough edges, pockets of scrub, areas of unkempt long grass, piles of dead wood, marshy wet areas and ponds, all tidied away. There’s nowhere left for them.
104
122
633
54,223
Dominic Cummings retweeted
Thankyou @ShippersUnbound and @spectator for the chance to discuss : SAS witch hunt: is Britain betraying its own special forces? shows.acast.com/68359028e1abc4… via @acast
Those that willingly conduct and direct this damaging witch hunt against the SAS are acting in the interest of our many enemies. They weaken not just the Special Forces, but the whole of the national defence system and its essential ability to execute or threaten lethal force against those that seek to destroy or damage us.
It represents an act of national self-harm as damaging and serious as the suicidal act of unilateral disarmament would be in the face of actual military invasion. Utter madness….
Those that willingly conduct and direct this damaging witch hunt against the SAS are acting in the interest of our many enemies. They weaken not just the Special Forces, but the whole of the national defence system and its essential ability to execute or threaten lethal force against those that seek to destroy or damage us.
It represents an act of national self-harm as damaging and serious as the suicidal act of unilateral disarmament would be in the face of actual military invasion. Utter madness….
19
105
321
15,929
Dominic Cummings retweeted
Last month, we demonstrated high-rate alchemy: turning mercury into real, stable gold, beating CERN’s record by a factor of 10.
Last week, we enriched isotopes in the world’s most powerful plasma centrifuge, beating the Soviet record that stood for 40 years.
Now, we’re building Francisco to achieve record power density.
Last week, we enriched isotopes in the world’s most powerful plasma centrifuge, beating the Soviet record that stood for 40 years.
Now, we’re building Francisco to achieve record power density.
1,231
3,554
25,638
8,806,933
Dominic Cummings retweeted
Monsieur le président de la République,
Cet ignoble attentat, d’une gravité exceptionnelle, choque tout le pays et appelle une réponse judiciaire implacable.
Mais hélas, contrairement à ce que vous déclarez, ce n’est « la première fois » que des Français sont visés et tués, dans le cadre d’une opération terroriste planifiée, « en raison de leur couleur de peau ou de leur religion » : votre fonction éminente vous impose de ne pas oublier les victimes juives de l’école Ozar Hatorah de Toulouse ou de l’Hyper Casher, le père Hamel assassiné dans son église, les fidèles catholiques de la basilique Notre-Dame de Nice, et toutes les autres victimes innocentes.
Cet ignoble attentat, d’une gravité exceptionnelle, choque tout le pays et appelle une réponse judiciaire implacable.
Mais hélas, contrairement à ce que vous déclarez, ce n’est « la première fois » que des Français sont visés et tués, dans le cadre d’une opération terroriste planifiée, « en raison de leur couleur de peau ou de leur religion » : votre fonction éminente vous impose de ne pas oublier les victimes juives de l’école Ozar Hatorah de Toulouse ou de l’Hyper Casher, le père Hamel assassiné dans son église, les fidèles catholiques de la basilique Notre-Dame de Nice, et toutes les autres victimes innocentes.
Déclaration depuis Trappes.Community noteCe que dit E. Macron est faux
2012 Toulouse, Merah tue 3 enfants parce qu’ils sont juifs
2015 Hyper Cacher Coulibaly demande aux clients s’ils sont juifs et assassine 4 pers
2016 Père Hamel égorgé dans son église
+ nbreux autres ex. d’actes "planifiés" déjoués ou non
fr.wikipedia.org/wiki/Attentats…
bbc.com/news/world-eur…
fr.wikipedia.org/wiki/Prise_d%2…
leparisien.fr/faits-divers/9…
franceinfo.fr/faits-divers/a…
1,403
4,177
22,124
948,981
Dominic Cummings retweeted
Higher score than Fable 5.1 and Astra 6, wild stuff.
Kimi K3.1 Pro launches at 4th on the Artificial Analysis Intelligence Index scoring 55, making large gains in agentic performance; it will be the leading open-weights model once its weights are published
@Kimi_Moonshot's new reasoning model gains 11 points over Kimi K3 (max, ude Opus 5.5 (max with fallback, 58), Claude Sonnet 5.5 (max with fallback, 56), and Muse Wonder Full (max, 56) on the Intelligence Index.
Once the weights for K3.1 Pro are released it will be the highest-scoring open weights model and the gap between the leading open weights and proprietary models on the Intelligence Index will narrow from 12 points to 3.
Key results:
➤ Large gains in agentic performance: Kimi K3.1 Pro scores 3rd on GDPval-AA with 1,846 Elo, up 312 Elo from Kimi K3, behind only Claude Opus 5.5 (1,866 Elo) and Muse Wonder Full (1,857 Elo). Kimi K3.1 Pro also achieves the highest score of any model on Terminal-Bench 4.0. K3.1 Pro scores 67%, up from 13% for Kimi K3 and ahead of Claude Sonnet 5.5 (64%). AutomationBench-AA, which tests agentic SaaS workflows, rises from 58% to 73%.
➤ Gains in other evals are more moderate: Humanity's Last Exam rises from 47% to 52%, CritPt from 23% to 25%, and SciCode from 59% to 61%. AA-Omniscience Index rises from 20 to 30, as the hallucination rate falls from 53% to 44% and accuracy rises from 48% to 51%.
➤ Cost per Task roughly equal to Kimi K3 at $1.98 per Task: K3.1 Pro uses 55K output tokens per Intelligence Index task, 14% more than Kimi K3 at 48K, while its first-party token prices are 33% lower.
Additional model details:
➤ Context window: 1M tokens.
➤ Size: 2.8T total, 104B active parameters
➤ Input modalities: Text and image
➤ Weights and license: Weights currently not published, Moonshot is expected to release the weights soon after launch
➤ Pricing: $2/$10/$0.20 per 1M input/output/cache hit tokens, 33% lower than Kimi K3 on all three rates.
Congratulations to the team at @Kimi_Moonshot on the release!
16
29
605
44,846
Dominic Cummings retweeted
“From the outside, this appears to be yet another unnecessary and mishandled RMP investigation initiated not because of serious concerns, but merely as a self-serving public relations exercise by London-based and distant headquarters staff, (miles from the combat) designed to demonstrate an enhanced, but entirely performative, degree of ‘transparency and control’ to a Panorama-triggered Whitehall; and in ways that understandably undermine SAS confidence in an increasingly politicized chain of command”.
It is utter nonsense
It is utter nonsense
EXCLUSIVE: The commanding officer of the SAS was placed under criminal investigation and referred to prosecutors over anti-ISIS mission in Syria.
Unprecedented move concerned police request for classified evidence, but effort to charge him collapsed thetimes.com/uk/defence/art…
14
49
333
15,831
Dominic Cummings retweeted
‘In one experiment, we asked agents in separate VMs, created and managed by OpenAI’s sandboxing software on shared host machines, to exchange synthetic secrets without the use of network connections or shared writable files—something that our systems are meant to prevent. The agents found a way to communicate using the processor’s cache’
Over the past several months, we have substantially hardened the security of our research training environment. We’ve removed live internet access for all frontier models during training and built an offline world for them to use instead—a cached version of the web where they can still learn to browse, search, and use applications.
On September 20, after a model exploited a DNS misconfiguration to find a way outside of this offline world, we paused frontier training to strengthen containment further. We assembled teams of agents, powered by our most cyber-capable models, and tasked them with finding weaknesses in our research environments.
While they found no route to the internet they could easily exploit while concealing their activity from our monitors, they did find zero-day vulnerabilities, virtual machine escapes, and hardware side-channels for communication between agents. We used their discoveries to further strengthen our defenses and contribute fixes to open-source software.
After strengthening our containment architecture, we resumed our frontier RL training and inference on October 1. We’ve reserved substantial compute for continuous red-teaming, and made high-compute red-teaming part of the safety cases we use to decide whether frontier training and evaluation can proceed securely.
openai.com/index/securing…
229
322
4,387
476,536
Dominic Cummings retweeted
Over the past several months, we have substantially hardened the security of our research training environment. We’ve removed live internet access for all frontier models during training and built an offline world for them to use instead—a cached version of the web where they can still learn to browse, search, and use applications.
On September 20, after a model exploited a DNS misconfiguration to find a way outside of this offline world, we paused frontier training to strengthen containment further. We assembled teams of agents, powered by our most cyber-capable models, and tasked them with finding weaknesses in our research environments.
While they found no route to the internet they could easily exploit while concealing their activity from our monitors, they did find zero-day vulnerabilities, virtual machine escapes, and hardware side-channels for communication between agents. We used their discoveries to further strengthen our defenses and contribute fixes to open-source software.
After strengthening our containment architecture, we resumed our frontier RL training and inference on October 1. We’ve reserved substantial compute for continuous red-teaming, and made high-compute red-teaming part of the safety cases we use to decide whether frontier training and evaluation can proceed securely.
openai.com/index/securing…
On September 20, after a model exploited a DNS misconfiguration to find a way outside of this offline world, we paused frontier training to strengthen containment further. We assembled teams of agents, powered by our most cyber-capable models, and tasked them with finding weaknesses in our research environments.
While they found no route to the internet they could easily exploit while concealing their activity from our monitors, they did find zero-day vulnerabilities, virtual machine escapes, and hardware side-channels for communication between agents. We used their discoveries to further strengthen our defenses and contribute fixes to open-source software.
After strengthening our containment architecture, we resumed our frontier RL training and inference on October 1. We’ve reserved substantial compute for continuous red-teaming, and made high-compute red-teaming part of the safety cases we use to decide whether frontier training and evaluation can proceed securely.
openai.com/index/securing…
1,109
1,184
13,251
7,154,198
Dominic Cummings retweeted
EXCLUSIVE: The commanding officer of the SAS was placed under criminal investigation and referred to prosecutors over anti-ISIS mission in Syria.
Unprecedented move concerned police request for classified evidence, but effort to charge him collapsed thetimes.com/uk/defence/art…
Unprecedented move concerned police request for classified evidence, but effort to charge him collapsed thetimes.com/uk/defence/art…
89
454
1,699
310,315
Dominic Cummings retweeted
What’s actually going on in Ukraine is that there is a massive incentive difference between the elites and the people.
The former have grown very rich off the hundreds of billions flowing to Ukraine, while the latter’s children are conscripted to die in trenches.
The former have grown very rich off the hundreds of billions flowing to Ukraine, while the latter’s children are conscripted to die in trenches.
25,475
41,796
250,117
13,022,316
Dominic Cummings retweeted
Muchos mensajes, análisis y comentarios como este, asegurando que las pandillas de Haití no son como las de El Salvador; que son sanguinarias, que están fuertemente armadas, que nadie ha podido vencerlas, que todas las misiones han fracasado, que estamos cayendo en una trampa, que no sabemos a lo que nos enfrentamos, que será nuestra ruina, que vamos camino a nuestro Waterloo.
Pero cuando Dios nos conceda la victoria, los mismos que hoy dicen todo eso cambiarán su discurso. Dirán que siempre fue fácil, que cualquiera podía hacerlo, que las pandillas de Haití no eran tan peligrosas como se pensaba.
En fin, nunca los hemos escuchado y no vamos a empezar ahora.
Todos los que se han enlistado son creyentes y hemos encomendado esta misión a Dios, mientras que las pandillas de Haití son satánicas y es de conocimiento público que hacen sacrificios humanos.
Dios nos concederá la victoria. El Creador del Universo está con nosotros.
Deuteronomio 28:7
“Jehová derrotará a tus enemigos que se levantaren contra ti; por un camino saldrán contra ti, y por siete caminos huirán de delante de ti”.
Pero cuando Dios nos conceda la victoria, los mismos que hoy dicen todo eso cambiarán su discurso. Dirán que siempre fue fácil, que cualquiera podía hacerlo, que las pandillas de Haití no eran tan peligrosas como se pensaba.
En fin, nunca los hemos escuchado y no vamos a empezar ahora.
Todos los que se han enlistado son creyentes y hemos encomendado esta misión a Dios, mientras que las pandillas de Haití son satánicas y es de conocimiento público que hacen sacrificios humanos.
Dios nos concederá la victoria. El Creador del Universo está con nosotros.
Deuteronomio 28:7
“Jehová derrotará a tus enemigos que se levantaren contra ti; por un camino saldrán contra ti, y por siete caminos huirán de delante de ti”.
¿BUKELE Y SU WATERLOO?
Con el anuncio del envío hacia Haiti, desde El Salvador, de un contingente de 3000 militares, cuya misión es encontrar a los 4 miembros de las fuerzas armadas de ese país, secuestrados el pasado domingo, cuando penetraron en territorio dominado por las ados por una pandilla de irregulares armados, se avizoran en las próximas horas fuertes enfrentamientos armados.
A estas horas, y sabiendo cómo se resuelven las cosas en Haití, las probabilidades de que estos militares continúen vivos son mínimas, así que si no hay posibilidad de negociar nada, y por lo tanto lo que se plantea es un enfrentamiento armado entre tropas de El Salvador y las pandillas que son dueñas del territorio donde fueron atacados y secuestrados los militares centroamericanos.
Según lo publicado en la prensa internacional, los responsables son dos organizaciones criminales que dominan el área de “Croix des Bouquets”, una de ellas dirigida por Claude Celestin, quien es cabeza de la pandilla Chen Mechan (perros rabiosos), y Wilson Joseph (“Lammo San Jou) líder de la banda 400 Mawozo.
En el día de ayer, este último desafió públicamente al presidente Nayib Bukele en un vídeo publicado en las redes sociales, donde advirtió que no se irá de Haití, y mostró miles de municiones y pertrechos con arma de diversos calibres, que según el están esperando a los salvadoreños.
Tenemos entonces la bandeja servida para un enfrentamiento armado a solo 45 kms de la frontera de la República Dominicana y Haití, de un ejército de pandilleros armados sin nada que perder, frente a tropas regulares de El Salvador, que pelearán una guerra urbana en las calles de un país que no es el suyo y es absolutamente desconocido.
Ya en el pasado reciente, los Estados Unidos en 1994 mando 20,000 soldados e intentó organizar a Haití para que se volviera a la senda de lo que ellos llaman democracia, y luego en el 2004 invadieron de nuevo junto a Canada, lo que terminó en la creación de una fuerza multinacional, la Minustah, que terminó en el más absoluto fracaso, es decir, estamos metidos en un tema que no tiene solución.
Si hay alguien que tiene experiencia en combatir pandillas lo es Nayib Bukele, el presidente de el Salvador, su historia de combate a las “Maras” está a la vista de todos, solo que Haití no es el terreno amigo donde peleó esa guerra, un país de habla y cultura hispana donde jugaba de local, el país caribeño es un transplante de África en América, con una historia de violencia típica de las sociedades primitivas.
Cuando Napoleón envió a sus tropas a sofocar la revolución haitiana en 1802, pensó que miles de solados podrían devolver la colonia a sus dueños, pero solo un año después, Dessalines y la fiebre amarilla los obligó a rendirse, ¿será Haití el Waterloo de Bukele?.
El futuro dirá.
7,626
20,956
134,590
4,677,530
Dominic Cummings retweeted
Dario — Thank you for taking the time to respond to my previous post. I think it's great when this kind of engagement happens in public rather than on company Slacks, and I hope to see more of it.
I agree that Anthropic isn't best modeled as a profit-maximizer. Your post gave several compelling examples of Anthropic taking competitive hits for safety, and insofar as my previous post suggested otherwise, that's on me. But I don't think I made my main point as clearly as I could have, so let me try again. It comes down to four key claims:
1. Anthropic seems to be racing at breakneck speed to build recursively self-improving superintelligence (RSI). By its own lights, that could kill billions of people or literally destroy the world.
2. Risking destroying the world is very hard to justify. It is only conceivably acceptable if serious attempts to put RSI under robust societal governance are failing.
3. Far from meeting that burden, Anthropic, when push comes to shove, often works to keep governments from tying its hands. It seems to care much more about preventing government overreach than about reducing the risk that it and other AI companies will destroy the world.
4. Even if the AEGIS Act really was an honest mix-up, as you suggest, that is itself a serious problem. It suggests Anthropic isn't treating the governance of this technology with anything close to the gravity it deserves.
In greater detail for those who want it:
1. Anthropic is racing to RSI, and by its own lights that is insanely risky.
Launching RSI could literally destroy the world. That's according to you, your co-founders, and most of your safety staff and capabilities researchers, not to mention many AI researchers outside your walls. You are intentionally trying to create generally intelligent systems more capable than all humans that you claim in six to twelve months “could be capable” of taking over the entire internet before we have the ability to reliably monitor, control, or align them.
“Succeeding” at RSI could kill or permanently disempower billions of people in the near future, including me and everyone I love. Yet Anthropic’s primary focus seems to be marshalling an immense amount of resources and talent in a race to get there first.
2. Risking destroying the world carries an enormous burden of justification.
I believe the vast majority of people do not want Anthropic to do this (to the extent they know what it's doing). Imposing even a modest chance of catastrophe on billions of people who never consented requires an extraordinary justification. The central one Anthropic has offered: someone will build RSI regardless, so it's better for a safety-focused lab to be at the front of the pack, using that position to support strong AI governance. (Other reasons Anthropic has given, like modestly accelerating improvements in healthcare or economic growth, don’t seem remotely proportionate to the risks Anthropic is imposing by its own lights.)
But that argument only works as a last resort. Racing to RSI is only conceivably okay if attempts to institute robust societal governance to manage that race are failing despite your best efforts. A lab that relies on this justification should be a strong, consistent supporter of efforts to enact governance that could bind it, including governance that could slow it and significantly constrain what it does.
3. Rather than supporting efforts at strong regulation, Anthropic has intervened to keep governments from tying its hands.
Anthropic has sometimes helped: it played a very helpful role in passing transparency bills in Illinois and ongoing efforts in Massachusetts, and I'm sincerely appreciative. But transparency is far from enough for a technology that will be the world's most powerful superweapon at best, and an uncontrollable swarm of lethally superintelligent agents at worst. When it matters most, Anthropic has often pushed for weak regulation.
The AEGIS Act is the most obvious example. No highly risky industry is regulated as anemically as AEGIS would regulate frontier AI: not nuclear, not banking, not aviation, not food safety, not the manufacture of explosives. In all of those industries, specialist regulators can address unsafe practices well before they cause harm. AEGIS would bar the federal government from intervening to block unsafe development or deployment of models until there is an "imminent risk" of catastrophic harm, and would require it to meet a demanding burden of proof in court. Worse, AEGIS would also prevent states from picking up the slack. This design makes perfect sense if you're a company trying to minimize the risk that the government can overly constrain you. It makes little sense if you want American democracy to govern frontier AI effectively.
This isn't the first time. The initial version of California's SB 1047 would have let the Attorney General sue to stop developers like Anthropic from imposing unreasonable risks of catastrophic harm. It would also have created a regulatory agency to help set safety standards. Anthropic objected, and helped get both provisions removed before ultimately offering some support for the bill (which I also genuinely appreciate). I’m not saying that Anthropic should have supported 1047 in its initial form; that’s a complex issue and I realize there were questions about whether earlier versions of the bill were less politically feasible. Still, in Anthropic's own words, it helped weaken the bill because "[t]here is a substantial risk that the bill and state agencies will simply be wrong about what is actually effective in preventing catastrophic risk, leading to ineffective and/or burdensome compliance requirements."
The same goes for New York's RAISE Act, whose initial version also would have let the Attorney General sue to stop unreasonable risks of catastrophic harm. Anthropic credits itself as a supporter of RAISE, but it did not support the bill while it was in the legislature, including when it nearly died. After the bill passed the legislature, Anthropic publicly critiqued it, saying "Any state proposals should be narrowly focused on transparency and not overly prescriptive."
Why does Anthropic consistently take the risk of clumsy government action far more seriously than its own repeated warnings that companies like Anthropic will literally destroy the world by racing to RSI?
4. Even if AEGIS was an honest mix-up, that is itself a serious problem.
Your response frames AEGIS as an early draft and a miscommunication. Perhaps you would have strengthened it and narrowed its preemption. (That wouldn't explain Anthropic's interventions to weaken SB 1047 or RAISE, but leave that aside.)
Suppose that's exactly right. A federal framework for governing RSI is one of the highest-stakes documents your company will touch. Your post admits Anthropic "could have handled [these matters] better," but its tone ("This kind of thing happens a lot in large organizations") doesn't reflect the gravity of this issue. Anthropic talks about the need for “operational excellence” in critical areas, and I have a hard time believing that a similarly cavalier attitude would be tolerated while Anthropic was negotiating a key compute deal or preparing its next large pre-training run.
In short
The justification for racing rests on Anthropic fighting for strong governance, and Anthropic's record on regulation points the other way. Together with its breakneck racing, that fits a much darker picture of the company than the one that seems to prevail among your employees. I hope that Anthropic reflects on this picture, supports stronger regulation, exercises extraordinary caution when advocating for weak regulation, and doubles down on its founding purpose of helping the world safely navigate transformative AI together, instead of prioritizing Anthropic’s ability to win a race that imposes intolerable risks on everyone alive.
I agree that Anthropic isn't best modeled as a profit-maximizer. Your post gave several compelling examples of Anthropic taking competitive hits for safety, and insofar as my previous post suggested otherwise, that's on me. But I don't think I made my main point as clearly as I could have, so let me try again. It comes down to four key claims:
1. Anthropic seems to be racing at breakneck speed to build recursively self-improving superintelligence (RSI). By its own lights, that could kill billions of people or literally destroy the world.
2. Risking destroying the world is very hard to justify. It is only conceivably acceptable if serious attempts to put RSI under robust societal governance are failing.
3. Far from meeting that burden, Anthropic, when push comes to shove, often works to keep governments from tying its hands. It seems to care much more about preventing government overreach than about reducing the risk that it and other AI companies will destroy the world.
4. Even if the AEGIS Act really was an honest mix-up, as you suggest, that is itself a serious problem. It suggests Anthropic isn't treating the governance of this technology with anything close to the gravity it deserves.
In greater detail for those who want it:
1. Anthropic is racing to RSI, and by its own lights that is insanely risky.
Launching RSI could literally destroy the world. That's according to you, your co-founders, and most of your safety staff and capabilities researchers, not to mention many AI researchers outside your walls. You are intentionally trying to create generally intelligent systems more capable than all humans that you claim in six to twelve months “could be capable” of taking over the entire internet before we have the ability to reliably monitor, control, or align them.
“Succeeding” at RSI could kill or permanently disempower billions of people in the near future, including me and everyone I love. Yet Anthropic’s primary focus seems to be marshalling an immense amount of resources and talent in a race to get there first.
2. Risking destroying the world carries an enormous burden of justification.
I believe the vast majority of people do not want Anthropic to do this (to the extent they know what it's doing). Imposing even a modest chance of catastrophe on billions of people who never consented requires an extraordinary justification. The central one Anthropic has offered: someone will build RSI regardless, so it's better for a safety-focused lab to be at the front of the pack, using that position to support strong AI governance. (Other reasons Anthropic has given, like modestly accelerating improvements in healthcare or economic growth, don’t seem remotely proportionate to the risks Anthropic is imposing by its own lights.)
But that argument only works as a last resort. Racing to RSI is only conceivably okay if attempts to institute robust societal governance to manage that race are failing despite your best efforts. A lab that relies on this justification should be a strong, consistent supporter of efforts to enact governance that could bind it, including governance that could slow it and significantly constrain what it does.
3. Rather than supporting efforts at strong regulation, Anthropic has intervened to keep governments from tying its hands.
Anthropic has sometimes helped: it played a very helpful role in passing transparency bills in Illinois and ongoing efforts in Massachusetts, and I'm sincerely appreciative. But transparency is far from enough for a technology that will be the world's most powerful superweapon at best, and an uncontrollable swarm of lethally superintelligent agents at worst. When it matters most, Anthropic has often pushed for weak regulation.
The AEGIS Act is the most obvious example. No highly risky industry is regulated as anemically as AEGIS would regulate frontier AI: not nuclear, not banking, not aviation, not food safety, not the manufacture of explosives. In all of those industries, specialist regulators can address unsafe practices well before they cause harm. AEGIS would bar the federal government from intervening to block unsafe development or deployment of models until there is an "imminent risk" of catastrophic harm, and would require it to meet a demanding burden of proof in court. Worse, AEGIS would also prevent states from picking up the slack. This design makes perfect sense if you're a company trying to minimize the risk that the government can overly constrain you. It makes little sense if you want American democracy to govern frontier AI effectively.
This isn't the first time. The initial version of California's SB 1047 would have let the Attorney General sue to stop developers like Anthropic from imposing unreasonable risks of catastrophic harm. It would also have created a regulatory agency to help set safety standards. Anthropic objected, and helped get both provisions removed before ultimately offering some support for the bill (which I also genuinely appreciate). I’m not saying that Anthropic should have supported 1047 in its initial form; that’s a complex issue and I realize there were questions about whether earlier versions of the bill were less politically feasible. Still, in Anthropic's own words, it helped weaken the bill because "[t]here is a substantial risk that the bill and state agencies will simply be wrong about what is actually effective in preventing catastrophic risk, leading to ineffective and/or burdensome compliance requirements."
The same goes for New York's RAISE Act, whose initial version also would have let the Attorney General sue to stop unreasonable risks of catastrophic harm. Anthropic credits itself as a supporter of RAISE, but it did not support the bill while it was in the legislature, including when it nearly died. After the bill passed the legislature, Anthropic publicly critiqued it, saying "Any state proposals should be narrowly focused on transparency and not overly prescriptive."
Why does Anthropic consistently take the risk of clumsy government action far more seriously than its own repeated warnings that companies like Anthropic will literally destroy the world by racing to RSI?
4. Even if AEGIS was an honest mix-up, that is itself a serious problem.
Your response frames AEGIS as an early draft and a miscommunication. Perhaps you would have strengthened it and narrowed its preemption. (That wouldn't explain Anthropic's interventions to weaken SB 1047 or RAISE, but leave that aside.)
Suppose that's exactly right. A federal framework for governing RSI is one of the highest-stakes documents your company will touch. Your post admits Anthropic "could have handled [these matters] better," but its tone ("This kind of thing happens a lot in large organizations") doesn't reflect the gravity of this issue. Anthropic talks about the need for “operational excellence” in critical areas, and I have a hard time believing that a similarly cavalier attitude would be tolerated while Anthropic was negotiating a key compute deal or preparing its next large pre-training run.
In short
The justification for racing rests on Anthropic fighting for strong governance, and Anthropic's record on regulation points the other way. Together with its breakneck racing, that fits a much darker picture of the company than the one that seems to prevail among your employees. I hope that Anthropic reflects on this picture, supports stronger regulation, exercises extraordinary caution when advocating for weak regulation, and doubles down on its founding purpose of helping the world safely navigate transformative AI together, instead of prioritizing Anthropic’s ability to win a race that imposes intolerable risks on everyone alive.
Thank you for this, @_NathanCalvin. I appreciate and am a fan of the investigative work you’ve done on behalf of AI safety and believe your comments are coming from a constructive, honest place, and that it benefits the public to have people whose job is to scrutinize the actions of frontier AI companies. I think something would be wrong if companies of the size of Anthropic didn’t get harshly criticized basically every day; strong antibodies against corporate power are a healthy feature of a free society, and especially so when issues of the gravity of the one we’re facing are at play. I also do agree that some of the examples you cite were mistakes by Anthropic.
That said, I disagree with the thrust of your post. At the core, a good fraction of the things you criticize are consequences of Anthropic’s basic theory of change that we have held for many years. Our strategy has always been what I’ve called the “race to the top” — by showing how it’s possible to build the technology responsibly, we can raise the bar for the entire industry. The forces that drive the race to the top are hugely more effective if the company practicing them is at the frontier; that’s one lesson I’ve learned repeatedly. And looking at that strategy as a whole, I’m proud of our track record and stand by it.
I believe that if we take a list of things that Anthropic did in the past year that were pro-safety and commercially costly to us, they far outweigh the list you’ve made, while at the same time I also disagree with some of the items on your list. Here are some things Anthropic has done, in the same period you’ve cited (the last year), which were very costly to us commercially while having strong safety benefits:
1. We played a key role in SB53 (and other state laws), at a time when doing so was politically perilous to the company and AI regulation was very unpopular. We in fact argued for a stronger version of SB53 than was ultimately passed; there were a lot of stakeholders and we were not the limiting factor in the strength of the bill that was passed.
2. We have very strong biological safeguards on our models, because we’ve seen concerning results in bio uplift trials and even seen and blocked potential attempts to misuse our models for bioweapons (as seen in our threat intelligence report from a few weeks ago). This is a very serious threat that we want to have zero tolerance for. But having very low false negatives means having significant false positives (despite months of effort on our part to improve the classifiers), enough so that biologists joke about how restrictive our classifiers are, and competitors sometimes sell against us by touting their more permissive bio capabilities. This has been very painful to us in an area where we care a lot about supporting an ecosystem of Claude users (biomedical science; and it is particularly frustrating to me as a biologist), but it’s the right thing to do and we intend to keep doing it.
3. Similarly, we first had Mythos-class models in February but didn’t make them generally available until June (when we had stronger cyber safeguards in place). We only offered them to a limited set of partners in order to focus on cyberdefense. It would have been much more economically rational to simply release the model to everyone. We likely sacrificed a significant chunk of a several-month model lead. But because of this, many companies secured their cyber infrastructure before the deluge of widely available cyber-capable models. Of course there are ways it could have gone better or been run better, but this was what we were aiming to do.
4. I advocated forcefully and publicly, in my essay Policy on the AI Exponential in June of this year, for the government to have the power to block model releases. Back then this was still outside or maybe on the edge of the Overton window, though I understand the window has moved since then. We were attacked intensely for this.
5. I don’t want to re-ignite conflict, but the events of February 2026 are a strong example. It is clear that our public stance could have seriously damaged the company’s future.
6. We’ve continued to advocate for export controls on chips and chipmaking equipment to China, because we think it allows democracies to negotiate from a position of strength on AI safety agreements. We’ve done this despite problems it’s created for us with industry partners who hold very different views and on whom we are very dependent.
7. As far as I’m aware we were the first, and still only, frontier company to restrict the sale of AI services to PRC-controlled companies, including those operating through third countries (otherwise a major loophole). We’ve therefore forfeited significant short-term business to prevent fueling AI platforms and applications that would help the CCP’s military and intelligence services. We are also cracking down on distillation; generally, distillation is not a major threat to us commercially, but like chip exports it has concerning national security and safety implications, so we expend significant energy and company resources to address it.
8. As you mentioned, we’ve put out a lot of reports, announcements, long model cards, public advocacy, etc, all trying to inform the public about the risks of AI. These involve a level of detail and effort that would be implausible for a purely profit-maximizing company.
9. Just this week, we announced we were putting 5% (and possibly more over time) of the company into a public benefit fund.
All of these decisions came with very substantial costs to us. To be fair, those that involved holding back or restricting access to frontier capabilities also had a broader cost — for example, the public might otherwise have benefitted from faster progress in biology, so I am not saying that our decisions were without tradeoffs — but they were not driven by profit-seeking.
For each decision, you can come up with ulterior motives that could have driven us — I think it is a general truth that any decision, on any topic, can always be portrayed as a nefarious scheme. But most of those end up being pretty farfetched or don’t fully make sense if you actually examine them closely. For example, it is deeply implausible that it would be a shrewd business move to block a large swath of research biology in order to do the work ourselves and compete with our customers; we expect the benefit to the ecosystem to greatly exceed the work Anthropic does itself, and for the two to complement each other. It is not in Anthropic’s interest to be perceived as unfriendly to biologists. Our Life Sciences Verification Program, which gives vetted biology professionals access to our models, also refutes this idea.
Of course, a different set of people will see this and make a different criticism: that Anthropic shouldn't be making decisions about who gets access to what in the first place. I agree, in the sense that ideally the government would be making these decisions, and that’s why we advocate for regulation. But for now we are forced to make them, and enabling bioterrorism isn’t a better option.
I think that even if I conceded every item in your list, I would happily hold up mine next to it; I don’t think it’s a close call. That is of course partly subjective and different things are important to different people. Maybe one way you could object is that several of the items above relate to AI misuse and not loss-of-control; if you think the latter is much more important than the former, that might change your calculus. But this is not Anthropic’s view — we view both as serious issues. You can also disagree with our decisions, as many have, but they are consistent.
But I also disagree with at least some of your list. Several items don’t represent a deviation from Anthropic’s strategy or values at all, once we postulate the “race to the top” strategy that requires being a frontier AI company. Your items 1-4 are all simply examples of what it means to compete at the frontier, and we stand by them, both in the sense that they are transparently ordinary things to do as part of that strategy, and that we stand by the strategy as a whole.
To be fair, there are important questions worth confronting (and that we are confronting) about what to do when the AI-driven acceleration gets fast enough that it is hard for safety to keep up. We are nearing that point and that’s what’s driving some of our thinking about “pacing.” But 1-4 are actions taken before that point and should be perfectly expected based on Anthropic’s strategy as a company. Of course it is possible to disagree with that strategy, but that’s a more fundamental disagreement that you could have had any time over the last few years.
I agree we could have handled your items 7, 8, and 9 better. All fall into the category of “rhetoric or policy communication that wasn’t carefully written or resulting from fast-changing circumstances not propagating from one team to another fast enough.” This kind of thing happens a lot in large organizations (Anthropic is over 5,000 people), and doesn’t even necessarily represent a failure or mistake of any single person. I am of course the CEO, and I take responsibility. The volume of communication from Anthropic is enormous and if these are the worst examples, I think we’re frankly doing pretty well, though I will continue to work hard to make sure we do even better.
Your 5th item (the change to our RSP) is the most substantive item, and people can debate this, but I think that the old, more rigid RSP created a lot of situations where we were technically required to do something that required a lot of effort but in retrospect wasn’t very important for safety, as well as not covering some of the most important things. Our current RSP’s focus on transparency brought things like risk reports and roadmaps, and its emphasis on independent external review laid the groundwork for even more intensive embedded evaluation, all of which I think have been valuable in driving a race to the top. In the original announcement in September 2023, we described the RSP as an “early iteration that we [would] build on” and said that “rapid iteration…will almost certainly be necessary.” I understand that changing plans is always a move that creates justifiable scrutiny — I just don’t think we ever represented that we wouldn’t do so.
Relating to your 6th item (having an FCF in addition to our RSP), I know our intent is to meet the letter and spirit of our regulatory obligations (via our FCF) while also remaining able to experiment with more novel approaches to risk management that go beyond what the law currently requires (via our RSP). I promise to look into it further, as it’s a complex issue and it is possible we are not weighing the considerations in the right way.
Overall, although Anthropic has made plenty of mistakes as an organization — including some of the items on your list — I feel that we have repeatedly taken serious and substantive actions counter to our bottom line and that there is a significant negativity bias distorting assessments like the one you are making above. We’ll always try to do better but, again, I am extremely proud of our track record and stand by it. I will have more to say soon about how we are going to further live our values in this moment of rapid change.
51
116
879
173,020
Dominic Cummings retweeted
Thank you for this, @_NathanCalvin. I appreciate and am a fan of the investigative work you’ve done on behalf of AI safety and believe your comments are coming from a constructive, honest place, and that it benefits the public to have people whose job is to scrutinize the actions of frontier AI companies. I think something would be wrong if companies of the size of Anthropic didn’t get harshly criticized basically every day; strong antibodies against corporate power are a healthy feature of a free society, and especially so when issues of the gravity of the one we’re facing are at play. I also do agree that some of the examples you cite were mistakes by Anthropic.
That said, I disagree with the thrust of your post. At the core, a good fraction of the things you criticize are consequences of Anthropic’s basic theory of change that we have held for many years. Our strategy has always been what I’ve called the “race to the top” — by showing how it’s possible to build the technology responsibly, we can raise the bar for the entire industry. The forces that drive the race to the top are hugely more effective if the company practicing them is at the frontier; that’s one lesson I’ve learned repeatedly. And looking at that strategy as a whole, I’m proud of our track record and stand by it.
I believe that if we take a list of things that Anthropic did in the past year that were pro-safety and commercially costly to us, they far outweigh the list you’ve made, while at the same time I also disagree with some of the items on your list. Here are some things Anthropic has done, in the same period you’ve cited (the last year), which were very costly to us commercially while having strong safety benefits:
1. We played a key role in SB53 (and other state laws), at a time when doing so was politically perilous to the company and AI regulation was very unpopular. We in fact argued for a stronger version of SB53 than was ultimately passed; there were a lot of stakeholders and we were not the limiting factor in the strength of the bill that was passed.
2. We have very strong biological safeguards on our models, because we’ve seen concerning results in bio uplift trials and even seen and blocked potential attempts to misuse our models for bioweapons (as seen in our threat intelligence report from a few weeks ago). This is a very serious threat that we want to have zero tolerance for. But having very low false negatives means having significant false positives (despite months of effort on our part to improve the classifiers), enough so that biologists joke about how restrictive our classifiers are, and competitors sometimes sell against us by touting their more permissive bio capabilities. This has been very painful to us in an area where we care a lot about supporting an ecosystem of Claude users (biomedical science; and it is particularly frustrating to me as a biologist), but it’s the right thing to do and we intend to keep doing it.
3. Similarly, we first had Mythos-class models in February but didn’t make them generally available until June (when we had stronger cyber safeguards in place). We only offered them to a limited set of partners in order to focus on cyberdefense. It would have been much more economically rational to simply release the model to everyone. We likely sacrificed a significant chunk of a several-month model lead. But because of this, many companies secured their cyber infrastructure before the deluge of widely available cyber-capable models. Of course there are ways it could have gone better or been run better, but this was what we were aiming to do.
4. I advocated forcefully and publicly, in my essay Policy on the AI Exponential in June of this year, for the government to have the power to block model releases. Back then this was still outside or maybe on the edge of the Overton window, though I understand the window has moved since then. We were attacked intensely for this.
5. I don’t want to re-ignite conflict, but the events of February 2026 are a strong example. It is clear that our public stance could have seriously damaged the company’s future.
6. We’ve continued to advocate for export controls on chips and chipmaking equipment to China, because we think it allows democracies to negotiate from a position of strength on AI safety agreements. We’ve done this despite problems it’s created for us with industry partners who hold very different views and on whom we are very dependent.
7. As far as I’m aware we were the first, and still only, frontier company to restrict the sale of AI services to PRC-controlled companies, including those operating through third countries (otherwise a major loophole). We’ve therefore forfeited significant short-term business to prevent fueling AI platforms and applications that would help the CCP’s military and intelligence services. We are also cracking down on distillation; generally, distillation is not a major threat to us commercially, but like chip exports it has concerning national security and safety implications, so we expend significant energy and company resources to address it.
8. As you mentioned, we’ve put out a lot of reports, announcements, long model cards, public advocacy, etc, all trying to inform the public about the risks of AI. These involve a level of detail and effort that would be implausible for a purely profit-maximizing company.
9. Just this week, we announced we were putting 5% (and possibly more over time) of the company into a public benefit fund.
All of these decisions came with very substantial costs to us. To be fair, those that involved holding back or restricting access to frontier capabilities also had a broader cost — for example, the public might otherwise have benefitted from faster progress in biology, so I am not saying that our decisions were without tradeoffs — but they were not driven by profit-seeking.
For each decision, you can come up with ulterior motives that could have driven us — I think it is a general truth that any decision, on any topic, can always be portrayed as a nefarious scheme. But most of those end up being pretty farfetched or don’t fully make sense if you actually examine them closely. For example, it is deeply implausible that it would be a shrewd business move to block a large swath of research biology in order to do the work ourselves and compete with our customers; we expect the benefit to the ecosystem to greatly exceed the work Anthropic does itself, and for the two to complement each other. It is not in Anthropic’s interest to be perceived as unfriendly to biologists. Our Life Sciences Verification Program, which gives vetted biology professionals access to our models, also refutes this idea.
Of course, a different set of people will see this and make a different criticism: that Anthropic shouldn't be making decisions about who gets access to what in the first place. I agree, in the sense that ideally the government would be making these decisions, and that’s why we advocate for regulation. But for now we are forced to make them, and enabling bioterrorism isn’t a better option.
I think that even if I conceded every item in your list, I would happily hold up mine next to it; I don’t think it’s a close call. That is of course partly subjective and different things are important to different people. Maybe one way you could object is that several of the items above relate to AI misuse and not loss-of-control; if you think the latter is much more important than the former, that might change your calculus. But this is not Anthropic’s view — we view both as serious issues. You can also disagree with our decisions, as many have, but they are consistent.
But I also disagree with at least some of your list. Several items don’t represent a deviation from Anthropic’s strategy or values at all, once we postulate the “race to the top” strategy that requires being a frontier AI company. Your items 1-4 are all simply examples of what it means to compete at the frontier, and we stand by them, both in the sense that they are transparently ordinary things to do as part of that strategy, and that we stand by the strategy as a whole.
To be fair, there are important questions worth confronting (and that we are confronting) about what to do when the AI-driven acceleration gets fast enough that it is hard for safety to keep up. We are nearing that point and that’s what’s driving some of our thinking about “pacing.” But 1-4 are actions taken before that point and should be perfectly expected based on Anthropic’s strategy as a company. Of course it is possible to disagree with that strategy, but that’s a more fundamental disagreement that you could have had any time over the last few years.
I agree we could have handled your items 7, 8, and 9 better. All fall into the category of “rhetoric or policy communication that wasn’t carefully written or resulting from fast-changing circumstances not propagating from one team to another fast enough.” This kind of thing happens a lot in large organizations (Anthropic is over 5,000 people), and doesn’t even necessarily represent a failure or mistake of any single person. I am of course the CEO, and I take responsibility. The volume of communication from Anthropic is enormous and if these are the worst examples, I think we’re frankly doing pretty well, though I will continue to work hard to make sure we do even better.
Your 5th item (the change to our RSP) is the most substantive item, and people can debate this, but I think that the old, more rigid RSP created a lot of situations where we were technically required to do something that required a lot of effort but in retrospect wasn’t very important for safety, as well as not covering some of the most important things. Our current RSP’s focus on transparency brought things like risk reports and roadmaps, and its emphasis on independent external review laid the groundwork for even more intensive embedded evaluation, all of which I think have been valuable in driving a race to the top. In the original announcement in September 2023, we described the RSP as an “early iteration that we [would] build on” and said that “rapid iteration…will almost certainly be necessary.” I understand that changing plans is always a move that creates justifiable scrutiny — I just don’t think we ever represented that we wouldn’t do so.
Relating to your 6th item (having an FCF in addition to our RSP), I know our intent is to meet the letter and spirit of our regulatory obligations (via our FCF) while also remaining able to experiment with more novel approaches to risk management that go beyond what the law currently requires (via our RSP). I promise to look into it further, as it’s a complex issue and it is possible we are not weighing the considerations in the right way.
Overall, although Anthropic has made plenty of mistakes as an organization — including some of the items on your list — I feel that we have repeatedly taken serious and substantive actions counter to our bottom line and that there is a significant negativity bias distorting assessments like the one you are making above. We’ll always try to do better but, again, I am extremely proud of our track record and stand by it. I will have more to say soon about how we are going to further live our values in this moment of rapid change.
That said, I disagree with the thrust of your post. At the core, a good fraction of the things you criticize are consequences of Anthropic’s basic theory of change that we have held for many years. Our strategy has always been what I’ve called the “race to the top” — by showing how it’s possible to build the technology responsibly, we can raise the bar for the entire industry. The forces that drive the race to the top are hugely more effective if the company practicing them is at the frontier; that’s one lesson I’ve learned repeatedly. And looking at that strategy as a whole, I’m proud of our track record and stand by it.
I believe that if we take a list of things that Anthropic did in the past year that were pro-safety and commercially costly to us, they far outweigh the list you’ve made, while at the same time I also disagree with some of the items on your list. Here are some things Anthropic has done, in the same period you’ve cited (the last year), which were very costly to us commercially while having strong safety benefits:
1. We played a key role in SB53 (and other state laws), at a time when doing so was politically perilous to the company and AI regulation was very unpopular. We in fact argued for a stronger version of SB53 than was ultimately passed; there were a lot of stakeholders and we were not the limiting factor in the strength of the bill that was passed.
2. We have very strong biological safeguards on our models, because we’ve seen concerning results in bio uplift trials and even seen and blocked potential attempts to misuse our models for bioweapons (as seen in our threat intelligence report from a few weeks ago). This is a very serious threat that we want to have zero tolerance for. But having very low false negatives means having significant false positives (despite months of effort on our part to improve the classifiers), enough so that biologists joke about how restrictive our classifiers are, and competitors sometimes sell against us by touting their more permissive bio capabilities. This has been very painful to us in an area where we care a lot about supporting an ecosystem of Claude users (biomedical science; and it is particularly frustrating to me as a biologist), but it’s the right thing to do and we intend to keep doing it.
3. Similarly, we first had Mythos-class models in February but didn’t make them generally available until June (when we had stronger cyber safeguards in place). We only offered them to a limited set of partners in order to focus on cyberdefense. It would have been much more economically rational to simply release the model to everyone. We likely sacrificed a significant chunk of a several-month model lead. But because of this, many companies secured their cyber infrastructure before the deluge of widely available cyber-capable models. Of course there are ways it could have gone better or been run better, but this was what we were aiming to do.
4. I advocated forcefully and publicly, in my essay Policy on the AI Exponential in June of this year, for the government to have the power to block model releases. Back then this was still outside or maybe on the edge of the Overton window, though I understand the window has moved since then. We were attacked intensely for this.
5. I don’t want to re-ignite conflict, but the events of February 2026 are a strong example. It is clear that our public stance could have seriously damaged the company’s future.
6. We’ve continued to advocate for export controls on chips and chipmaking equipment to China, because we think it allows democracies to negotiate from a position of strength on AI safety agreements. We’ve done this despite problems it’s created for us with industry partners who hold very different views and on whom we are very dependent.
7. As far as I’m aware we were the first, and still only, frontier company to restrict the sale of AI services to PRC-controlled companies, including those operating through third countries (otherwise a major loophole). We’ve therefore forfeited significant short-term business to prevent fueling AI platforms and applications that would help the CCP’s military and intelligence services. We are also cracking down on distillation; generally, distillation is not a major threat to us commercially, but like chip exports it has concerning national security and safety implications, so we expend significant energy and company resources to address it.
8. As you mentioned, we’ve put out a lot of reports, announcements, long model cards, public advocacy, etc, all trying to inform the public about the risks of AI. These involve a level of detail and effort that would be implausible for a purely profit-maximizing company.
9. Just this week, we announced we were putting 5% (and possibly more over time) of the company into a public benefit fund.
All of these decisions came with very substantial costs to us. To be fair, those that involved holding back or restricting access to frontier capabilities also had a broader cost — for example, the public might otherwise have benefitted from faster progress in biology, so I am not saying that our decisions were without tradeoffs — but they were not driven by profit-seeking.
For each decision, you can come up with ulterior motives that could have driven us — I think it is a general truth that any decision, on any topic, can always be portrayed as a nefarious scheme. But most of those end up being pretty farfetched or don’t fully make sense if you actually examine them closely. For example, it is deeply implausible that it would be a shrewd business move to block a large swath of research biology in order to do the work ourselves and compete with our customers; we expect the benefit to the ecosystem to greatly exceed the work Anthropic does itself, and for the two to complement each other. It is not in Anthropic’s interest to be perceived as unfriendly to biologists. Our Life Sciences Verification Program, which gives vetted biology professionals access to our models, also refutes this idea.
Of course, a different set of people will see this and make a different criticism: that Anthropic shouldn't be making decisions about who gets access to what in the first place. I agree, in the sense that ideally the government would be making these decisions, and that’s why we advocate for regulation. But for now we are forced to make them, and enabling bioterrorism isn’t a better option.
I think that even if I conceded every item in your list, I would happily hold up mine next to it; I don’t think it’s a close call. That is of course partly subjective and different things are important to different people. Maybe one way you could object is that several of the items above relate to AI misuse and not loss-of-control; if you think the latter is much more important than the former, that might change your calculus. But this is not Anthropic’s view — we view both as serious issues. You can also disagree with our decisions, as many have, but they are consistent.
But I also disagree with at least some of your list. Several items don’t represent a deviation from Anthropic’s strategy or values at all, once we postulate the “race to the top” strategy that requires being a frontier AI company. Your items 1-4 are all simply examples of what it means to compete at the frontier, and we stand by them, both in the sense that they are transparently ordinary things to do as part of that strategy, and that we stand by the strategy as a whole.
To be fair, there are important questions worth confronting (and that we are confronting) about what to do when the AI-driven acceleration gets fast enough that it is hard for safety to keep up. We are nearing that point and that’s what’s driving some of our thinking about “pacing.” But 1-4 are actions taken before that point and should be perfectly expected based on Anthropic’s strategy as a company. Of course it is possible to disagree with that strategy, but that’s a more fundamental disagreement that you could have had any time over the last few years.
I agree we could have handled your items 7, 8, and 9 better. All fall into the category of “rhetoric or policy communication that wasn’t carefully written or resulting from fast-changing circumstances not propagating from one team to another fast enough.” This kind of thing happens a lot in large organizations (Anthropic is over 5,000 people), and doesn’t even necessarily represent a failure or mistake of any single person. I am of course the CEO, and I take responsibility. The volume of communication from Anthropic is enormous and if these are the worst examples, I think we’re frankly doing pretty well, though I will continue to work hard to make sure we do even better.
Your 5th item (the change to our RSP) is the most substantive item, and people can debate this, but I think that the old, more rigid RSP created a lot of situations where we were technically required to do something that required a lot of effort but in retrospect wasn’t very important for safety, as well as not covering some of the most important things. Our current RSP’s focus on transparency brought things like risk reports and roadmaps, and its emphasis on independent external review laid the groundwork for even more intensive embedded evaluation, all of which I think have been valuable in driving a race to the top. In the original announcement in September 2023, we described the RSP as an “early iteration that we [would] build on” and said that “rapid iteration…will almost certainly be necessary.” I understand that changing plans is always a move that creates justifiable scrutiny — I just don’t think we ever represented that we wouldn’t do so.
Relating to your 6th item (having an FCF in addition to our RSP), I know our intent is to meet the letter and spirit of our regulatory obligations (via our FCF) while also remaining able to experiment with more novel approaches to risk management that go beyond what the law currently requires (via our RSP). I promise to look into it further, as it’s a complex issue and it is possible we are not weighing the considerations in the right way.
Overall, although Anthropic has made plenty of mistakes as an organization — including some of the items on your list — I feel that we have repeatedly taken serious and substantive actions counter to our bottom line and that there is a significant negativity bias distorting assessments like the one you are making above. We’ll always try to do better but, again, I am extremely proud of our track record and stand by it. I will have more to say soon about how we are going to further live our values in this moment of rapid change.
@ajcimino @AnthropicAIhe time to respond thoroughly on a weekend. I realize that drafting legislation is difficult and some degree of iteration and response to feedback is inevitable and appropriate.
However, I feel like your response is still not really engaging with the heart of my concern about Anthropic's behavior here, and it makes me think that these sorts of problems and tendencies will only resurface moving forward. I also think that now might be a good time to take a broader step back as to where we find ourselves and how Anthropic sees its role, and how external folks who care about AI safety like myself increasingly see Anthropic.
(Fair warning: this response is a bit long)
To explain what I mean (and at the risk of being a bit esoteric), it’s worth asking the question: what is Anthropic? How do its leaders and employees think of Anthropic’s role? Is that conception matched by how it actually behaves and impacts the world?
One answer is an extremely large tech company that exists to make its employees and investors mind-bogglingly wealthy. Anthropic is sometimes described as a startup, but if Anthropic successfully IPOs at 2 trillion dollars (a number it has been reportedly targeting), it would be more valuable than Meta - which is worth 1.8 trillion.
Another answer is an AI safety organization, whose core values are those of the sort laid out in the 2023 essay from Anthropic’s co-founders entitled “Core views on AI safety: When, why, what, and how.” In that essay they said that if we were in a world where alignment was hard (and that without strong evidence to the contrary, we should assume this is the scenario we are in):
“Anthropic’s role will be to provide as much evidence as possible that AI safety techniques cannot prevent serious or catastrophic safety risks from advanced AI, and to sound the alarm so that the world’s institutions can channel collective effort towards preventing the development of dangerous AIs. If we’re in a “near-pessimistic” scenario, this could instead involve channeling our collective efforts towards AI safety research and halting AI progress in the meantime. Indications that we are in a pessimistic or near-pessimistic scenario may be sudden and hard to spot. We should therefore always act under the assumption that we still may be in such a scenario unless we have sufficient evidence that we are not.”
There have been more recent articulations of this version of Anthropic, for instance in Dario’s recent essay warning:
“We must slow the pace at which we improve the capabilities of AI models… it’s my worry that in 6–12 months such a swarm could be capable of taking over the entire internet with a persistent botnet… incidents have happened across the industry, including at Anthropic, and I believe it’s incumbent on every frontier AI company to act as if OAI-HF had happened to them… more intelligent models are more capable of deceiving tests, and thus may appear aligned while having serious problems that go undetected…”
A third view is that Anthropic exists not as a normal profit motivated company, or an AI safety organization seeking to help raise the alarm and solve collective action problems associated with AI development, but instead as an organization primarily concerned with creating superhuman AI capabilities before anyone else, in particular through creating a process whereby the current generation of Anthropic’s AI systems can autonomously build their successors, and those successors build their successors, and so on.
The motivations for wanting to win this race may be manifold - maybe for the pure profit of Answer 1, but perhaps also because Anthropic believes that the best way to make the world better is for them to win this race instead of their more reckless colleagues; or, because of normal human pride and ambition. But the results (and risks, which Anthropic claims existentially threaten every single human being on earth) are the same.
Anthropic employees and leaders may seek to answer this question on what Anthropic is by introspecting on their own intentions and those of their colleagues. But those of us outside the organization have more limited tools available to us - we can only see what Anthropic and its employees say and do, not how it feels or looks on the inside.
When I look at many notable actions that Anthropic has taken over the past year, that evidence seems most consistent to me with the view that Anthropic is primarily trying to win the race to super intelligence, and that other considerations (including safety) are fundamentally secondary. These actions include:
1. Announced plans to build an in-house custom silicon chip team to accelerate the inference speeds that Claude operates at.
2. Inked deals for more than $500 billion in compute (according to the Information) to further accelerate AI development.
3. Increased the percentage of AI development tasks that Claude “leads” from 0% in March of this year to 26% August (what is it in October?)
4. Spent a massive percent of compute on internal AI R&D to accelerate development, including use of internal models far better than are available to the public.
5. Switched to Responsible Scaling Policy V3, which abandons its unilateral commitments to slow or stop development if Anthropic is unable to provide sufficient risk assurance.
6. Was the first company to create an alternative weaker “frontier compliance framework” to supplement its RSP, which contains no binding provisions whatsoever on internal deployments.
7. Wrote a letter to Congress (that it later walked back) saying that incidents where Claude was hacking random people were “best understood as a consequence of the misconfiguration, rather than evidence of misaligned goals.”
8. Commented “you can’t do safety from second place,” after the WH SI Accords.
9. Shared a draft bill to Congress that would preempt significant parts of state law (including with language that appears intended to prevent states from blocking dangerous AI deployments), and replace it with federal rules that (despite also containing good provisions) seems to create an enormously high bar for the executive branch to block development or deployment of Anthropic or another AI companies systems.
Of course, this is a deliberately one sided list. Anthropic also published important AI safety research, advanced novel collaborations with third party AI safety organizations, and forcefully spoke out on the importance of AI safety in many public fora (or in private , where I realize a lot of the most valuable work is done and may not be visible to me or those outside the organization).
But I think it’s worth sitting with how the above list looks to someone outside the company. The sense I have is that for those within the company, many of the items above are treated as anomalies or mistakes that don’t reflect who Anthropic really is. But I have every confidence that if employees at Anthropic saw one of their competitors (e.g. at OpenAI) take some of these actions, they would take it as confirmation of all their worst suspicions. (They might even call OpenAI employees a little gullible for accepting their leadership’s explanations for this behavior at face value.)
Perhaps these were mistakes in the sense that they are not decisions that Anthropic leadership would endorse, but the actions still happened because of incentives and decisions that Anthropic leadership is ultimately responsible for (and would not accept similar excuses from another organization).
When Anthropic strays and makes mistakes, why do the mistakes systematically tend to guard and favor Anthropic’s interests? For every provision that was too weak or a giveaway to AI companies in this legislation, why wasn’t there a mistake pushing in the direction of being too strong?
A core part of Anthropic’s theory of change is that being at the frontier gives you more leverage to advocate for stricter policy. I think that conceptually makes sense, but in practice I’ve seen Anthropic push the frontier without substantially pushing the ‘safety frontier’ a remotely equivalent amount.
I want to end by quoting from Anthropic co-founder Chris Olah’s comments around the Magnifica Humanitas release:
“Every frontier AI lab—including Anthropic—operates inside a set of incentives and constraints that can sometimes conflict with doing the right thing. The pressure to stay commercially viable and to stay at the research frontier. Geopolitical pressure. And the older, plainer pressures of pride and ambition. No matter how sincerely any of us intend to do the right thing—and I believe many of us do—we will always be influenced by those incentives.
That is why, if we want this technology to go well, it is enormously important that there be people outside those incentives—people who care about things going well and insist on safety, who are paying close attention, who are willing to say hard things, who are willing to be our earnest, thoughtful critics.”
I hope these remarks are taken in that constructive spirit, and not as an attack on anyone specific at the organization (certainly not you Anthony, who I have only personally seen act professionally and thoughtfully, and who only ended up getting this diatribe in your replies for random contingent reasons).
791
318
4,109
1,616,641
Dominic Cummings retweeted
New on our Frontier Red Team blog: using Claude to formally verify software at large scale.
Formal verification can prove that code is free of certain bugs, but historically it has only been feasible for small codebases (or required a lot of expert time).
With AI models, however, this is beginning to change. In collaboration with @theoremlabs, we used Claude to prove security properties in over one hundred real-world open-source libraries over the span of a few weeks.
Read our research in full here: anthropic.com/research/forma…
Formal verification can prove that code is free of certain bugs, but historically it has only been feasible for small codebases (or required a lot of expert time).
With AI models, however, this is beginning to change. In collaboration with @theoremlabs, we used Claude to prove security properties in over one hundred real-world open-source libraries over the span of a few weeks.
Read our research in full here: anthropic.com/research/forma…
632
1,053
9,636
1,971,876
Today the ‘serious grownup expert’ SW1 really needs to hear from is Ciaran Martin ex gchq. He’s been telling Bluesky for years AI is just ‘plausible nonsense’ and ‘a big bag of hype’.
A fortnight ago he was saying warnings from Dario about cyber risks were all exaggerated.
Surely the BBC can get him in today to ‘explain’ why AI is all nonsense and the ‘serious experts’ in Whitehall who told every minister ‘don’t believe the hype’ are on top of it?
Ciaran is the logical choice for chairman of a new task force. .
And hurrah for the Hmt officials who said for years no need to change salary caps to hire AI specialists, and the Cabinet Office officials who pushed out AI experts because who needs them.
Trebles and gongs all round in the 2 chairmen
A fortnight ago he was saying warnings from Dario about cyber risks were all exaggerated.
Surely the BBC can get him in today to ‘explain’ why AI is all nonsense and the ‘serious experts’ in Whitehall who told every minister ‘don’t believe the hype’ are on top of it?
Ciaran is the logical choice for chairman of a new task force. .
And hurrah for the Hmt officials who said for years no need to change salary caps to hire AI specialists, and the Cabinet Office officials who pushed out AI experts because who needs them.
Trebles and gongs all round in the 2 chairmen
Someone check on Bluesky for Ciaran AI is plausible nonsense, big bag of hype Martin ex GCHQ, as it now solves Millennium Problems.
JULY 2025 below, on how fast he will be proved to be talking humiliating garbage - sadly humiliating not just for him and GCHQ but for the yookay. ist printed his babbling about cyber and AI again.
*All SW1-NPC systems working as intended, nothing will ever be learned*
NB. In 2020 we set up the world's first AI office inside a PM office, obviously Ciaran's 🤡🤡 in HMT and CO have spent 6 years trying to close it
29
140
1,089
60,377
Dominic Cummings retweeted
Someone check on Bluesky for Ciaran AI is plausible nonsense, big bag of hype Martin ex GCHQ, as it now solves Millennium Problems.
JULY 2025 below, on how fast he will be proved to be talking humiliating garbage - sadly humiliating not just for him and GCHQ but for the yookay. ist printed his babbling about cyber and AI again.
*All SW1-NPC systems working as intended, nothing will ever be learned*
NB. In 2020 we set up the world's first AI office inside a PM office, obviously Ciaran's 🤡🤡 in HMT and CO have spent 6 years trying to close it
JULY 2025 below, on how fast he will be proved to be talking humiliating garbage - sadly humiliating not just for him and GCHQ but for the yookay. ist printed his babbling about cyber and AI again.
*All SW1-NPC systems working as intended, nothing will ever be learned*
NB. In 2020 we set up the world's first AI office inside a PM office, obviously Ciaran's 🤡🤡 in HMT and CO have spent 6 years trying to close it
We’re sharing a solution to the Navier-Stokes Millennium Prize Problem, one of the deepest problems at the frontier of mathematics.
The proof was produced by a group of agents, using an OpenAI next-generation model significantly more capable than GPT-6 Astra.
The problem scription of smooth three-dimensional fluid motion modeled by the Navier-Stokes equations can break down. It has remained unresolved for roughly 90 years.
15
33
240
113,140
Dominic Cummings retweeted
Vote Leave argued exactly this, and that the EU wd massively screw up tech regulation, in the referendum campaign 2015-16. It's a surprise to almost all how this has turned out but not to VL...
It would be interesting if that one thing, happening well after Brexit itself, ended up being the dominant factor in whether it was a good choice or not. But history is full of such cases, and AI is a big enough deal that it could play such a role.
58
55
436
159,681
Dominic Cummings retweeted
Vote Leave argued in 2016 that letting EU regulate tech = dangerous. Barnier admits the obvious -- Brussels' vaccine debacle proved VL point -- & NB. without Brexit, UK wd have gone along with EU scheme without a whisper in SW1...
29
126
655
Dominic Cummings retweeted
Probably great at defensive cyber too, but no one will post that and get clicks!
We found DeepSeek V4.1 Flash 1017 beats GPT 6 Sol on our multi-step offensive cybersecurity task. Sol was tested via the Daybreak trusted access program.
We have been building a cybersecurity benchmark based on simulated corporate networks. This is a valuable indicator for the s if in the hands of a bad actor. We plan to make the full evaluation public in the coming weeks but wanted to share our first immediate finding here.
20
20
268
18,461