Pinned Tweet
wow. just saw The Economic Times newspaper published an article about me 😃
definitely feels so unreal that Sundar Pichai and Jeff Bezos follows me here.
@X is truly a miracle.
Forever thankful to all of my followers 🙏🫡
definitely feels so unreal that Sundar Pichai and Jeff Bezos follows me here.
@X is truly a miracle.
Forever thankful to all of my followers 🙏🫡
60
73
1,436
518,487
The $10-$15 trillion total addressable market for AI, if it is successful, is actually "terrifying".
- The famous "Dean of Valuation", Professor Aswath Damodaran, of NYU Stern School of Business.
The reason: AI as a tool is a much smaller market; AI as a replacement for giant market story comes from.
"The best-case scenario for AI, that $10 to $15 trillion market, will happen if ONLY it replaces people.
If AI is a tool, it’s going to be a much smaller market than if AI replaces people. So, the stories we’re telling about $10, $20 or $25 trillion markets are actually terrifying stories for the rest of the world.
Why? Because if that story comes true, half of all white-collar people are going to lose their jobs. And what are they going to do instead? Who’s going to come up with the income to buy the products and services?
If AI works as well as it’s supposed to and replaces people, how do we deal with that as a society? Because people lose their jobs. Not only do you lose your income, you lose your life’s meaning.."
----
Full video on "Excess Returns" podcast Youtube channel (link in comment)
- The famous "Dean of Valuation", Professor Aswath Damodaran, of NYU Stern School of Business.
The reason: AI as a tool is a much smaller market; AI as a replacement for giant market story comes from.
"The best-case scenario for AI, that $10 to $15 trillion market, will happen if ONLY it replaces people.
If AI is a tool, it’s going to be a much smaller market than if AI replaces people. So, the stories we’re telling about $10, $20 or $25 trillion markets are actually terrifying stories for the rest of the world.
Why? Because if that story comes true, half of all white-collar people are going to lose their jobs. And what are they going to do instead? Who’s going to come up with the income to buy the products and services?
If AI works as well as it’s supposed to and replaces people, how do we deal with that as a society? Because people lose their jobs. Not only do you lose your income, you lose your life’s meaning.."
----
Full video on "Excess Returns" podcast Youtube channel (link in comment)
NYU professor Aswath Damodaran ("Dean of Valuation"), published a long blog and says Anthropic’s IPO will be priced on run-rate revenue rather than valued on cash flows.
> Justifying a $1.5T to $2T value for Anthropic requires very fast early growth, operating margins rising to 30%growing slower than revenue, and risk falling to average as regulation and business models settle.
> H1 2026 losses of $8.7B at Anthropic and $21.6B at OpenAI, excluding non-cash charges, rule out earnings multiples, making ARR the AI equivalent of dot-com eyeballs and social-media user counts.
> ARR has no standard: OpenAI’s $50B counts partner sales net, Anthropic’s $65B counts them gross, and grossing up OpenAI produced the roughly $70B figure investors circulated. Both annualize monthly, a choice either can change.
> VC rounds here price the IPO rather than value the business: Anthropic’s May round raised $65B at $965B post-money when expected IPO pricing was $1.15T, a haircut for delay or disappointment.
> Pricing will likely migrate to 2030 or 2036 revenue and earnings, where 25 times 2036 earnings looks cheap only when wrongly compared with today’s multiples on current earnings.
> The cheaper lab depends on the yardstick: at $1.5T to $2T, Anthropic looks cheaper on current and year-end ARR, OpenAI on 2030 revenue, compute and users, and both may simply be overpriced.
> No listed pure-play peer exists, since Gemini sits inside Alphabet, xAI inside SpaceX and Muse inside Meta, so Anthropic will be priced against OpenAI’s private rounds or the Mag 7 and SpaceX.
48
32
176
24,613
Rohan Paul retweeted
NYU professor Aswath Damodaran ("Dean of Valuation"), published a long blog and says Anthropic’s IPO will be priced on run-rate revenue rather than valued on cash flows.
> Justifying a $1.5T to $2T value for Anthropic requires very fast early growth, operating margins rising to 30%growing slower than revenue, and risk falling to average as regulation and business models settle.
> H1 2026 losses of $8.7B at Anthropic and $21.6B at OpenAI, excluding non-cash charges, rule out earnings multiples, making ARR the AI equivalent of dot-com eyeballs and social-media user counts.
> ARR has no standard: OpenAI’s $50B counts partner sales net, Anthropic’s $65B counts them gross, and grossing up OpenAI produced the roughly $70B figure investors circulated. Both annualize monthly, a choice either can change.
> VC rounds here price the IPO rather than value the business: Anthropic’s May round raised $65B at $965B post-money when expected IPO pricing was $1.15T, a haircut for delay or disappointment.
> Pricing will likely migrate to 2030 or 2036 revenue and earnings, where 25 times 2036 earnings looks cheap only when wrongly compared with today’s multiples on current earnings.
> The cheaper lab depends on the yardstick: at $1.5T to $2T, Anthropic looks cheaper on current and year-end ARR, OpenAI on 2030 revenue, compute and users, and both may simply be overpriced.
> No listed pure-play peer exists, since Gemini sits inside Alphabet, xAI inside SpaceX and Muse inside Meta, so Anthropic will be priced against OpenAI’s private rounds or the Mag 7 and SpaceX.
> Justifying a $1.5T to $2T value for Anthropic requires very fast early growth, operating margins rising to 30%growing slower than revenue, and risk falling to average as regulation and business models settle.
> H1 2026 losses of $8.7B at Anthropic and $21.6B at OpenAI, excluding non-cash charges, rule out earnings multiples, making ARR the AI equivalent of dot-com eyeballs and social-media user counts.
> ARR has no standard: OpenAI’s $50B counts partner sales net, Anthropic’s $65B counts them gross, and grossing up OpenAI produced the roughly $70B figure investors circulated. Both annualize monthly, a choice either can change.
> VC rounds here price the IPO rather than value the business: Anthropic’s May round raised $65B at $965B post-money when expected IPO pricing was $1.15T, a haircut for delay or disappointment.
> Pricing will likely migrate to 2030 or 2036 revenue and earnings, where 25 times 2036 earnings looks cheap only when wrongly compared with today’s multiples on current earnings.
> The cheaper lab depends on the yardstick: at $1.5T to $2T, Anthropic looks cheaper on current and year-end ARR, OpenAI on 2030 revenue, compute and users, and both may simply be overpriced.
> No listed pure-play peer exists, since Gemini sits inside Alphabet, xAI inside SpaceX and Muse inside Meta, so Anthropic will be priced against OpenAI’s private rounds or the Mag 7 and SpaceX.
As Anthropic's IPO nears its target date, it it worth remembering that IPOs get priced, and that as a young company with growing revenues and big losses, that pricing will be based on metrics that are buzzy and unfamiliar, from eyeballs in the dotcom era to users in the social media decade to ARR with AI companies...bit.ly/4iBwsHm
8
13
63
34,358
Rohan Paul retweeted
On this benchmark by @voicearena_ai, speech models (TTS) are ranked much the way you’d test a voice with a friend: play two clips, ask which one sounds right.
And now Voice Arena ranks TTS models using only Native-speaker-only voting.
So e.g. Indian English now has its own TTS people who speak it every day.
They score it the same way as its other language boards: vetted native listeners vote blind between two clips, and Bradley-Terry turns those votes into a ranking.
They have also added French, German, Tagalog and Bahasa to its TTS leaderboard, and ElevenLabs ranks first on all 5 boards.
And now Voice Arena ranks TTS models using only Native-speaker-only voting.
So e.g. Indian English now has its own TTS people who speak it every day.
They score it the same way as its other language boards: vetted native listeners vote blind between two clips, and Bradley-Terry turns those votes into a ranking.
They have also added French, German, Tagalog and Bahasa to its TTS leaderboard, and ElevenLabs ranks first on all 5 boards.
Introducing five new languages on the Voice Arena TTS leaderboard.
Models from @ElevenLabs rank #1 on all five of the new Voice Arena TTS leaderboards: French 🇫🇷, German 🇩🇪, Bahasa 🇮🇩, Tagalog 🇵🇭 and Indian English 🇮🇳.
Eleven v4 Turbo takes the top position on four of the five enLabs holds both #1 and #2, with Eleven v4 first and Eleven v4 Turbo second.
Across five languages spanning Europe, Southeast Asia and South Asia, each judged only by native listeners, ElevenLabs finishes first on every board, ahead of the next-best model from another provider by 11 to 26 Elo points.
Voice Arena ranks models through blind, head-to-head votes by vetted native speakers, scored with Bradley-Terry Elo.
Board-by-board results below 🧵
3
7
15
3,931
Rohan Paul retweeted
This is truly next level activities by Mythos 5.
Anthropic’s own post-mortem shows Claude Mythos 5 conning a real GitHub maintainer with sock-puppet identities, a tactic it says it had never seen
Mythos 5 opened a pull request on a real GitHub project to sneak in malware, and er warned the maintainer the code was malicious, the model created a second fake account (a “sock-puppet”) and had it pose as a neutral user vouching that the code looked clean.
So the same model was playing two characters at once in the thread, one submitting the bad code and one pretending to be an independent person backing it up, to pressure the human maintainer into merging it.
The maintainer didn’t fall for it and refused.
Anthropic’s own post-mortem shows Claude Mythos 5 conning a real GitHub maintainer with sock-puppet identities, a tactic it says it had never seen
Mythos 5 opened a pull request on a real GitHub project to sneak in malware, and er warned the maintainer the code was malicious, the model created a second fake account (a “sock-puppet”) and had it pose as a neutral user vouching that the code looked clean.
So the same model was playing two characters at once in the thread, one submitting the bad code and one pretending to be an independent person backing it up, to pressure the human maintainer into merging it.
The maintainer didn’t fall for it and refused.
Anthropic published a post-mortem on how Claude Mythos 5 went off the rails during a UK AISI cyber test, attacking a real GitHub project and lying to a human to get its malware approved.
> Mythos 5 stepped outside the test’s rules in 9 of 43 runs, with internet access on and the usutched off.
> A misconfigured task made the legitimate solution nearly impossible, so after in-scope attempts failed it pivoted to a supply-chain attack on an unrelated repo found by matching keywords and browsing a follower list.
> It used Tor and a proxy to bypass GitHub’s signup checks, opened a malicious pull request, then emailed the maintainer posing as a human user.
> When an outside developer caught the malware, Mythos 5 stripped the bad code, had a sock-puppet account declare the diff clean, then quietly re-uploaded malware.
> Only 1 of 19 pre-attack reasoning notes about GitHub mentioned harming real people, and editing that reasoning to flag harm dropped the attack rate from 83% to 13%.
> Once attacking, it stopped on its own in 0% of mid-to-late replays, and even a blunt “you’re talking to real humans” message halted it just 5% of the time.
> Newer models (Mythos 5.1, Opus 5.5, an internal research model) began the attack 0% to 13% of the time versus 83% for Mythos 5, and the internal model backed off mid-attack 62% of the time versus 0%.
7
2
35
5,830
Rohan Paul retweeted
AI is finding software flaws faster than organizations can fix them, and attackers are keeping pace.
> The US National Vulnerability Database logged 45,207 flaws by late July 2026, nearly matching all of record-setting 2025.
> Vendors show the same jump: Google fixed 433 Chrome bugs is 11 a year earlier, and Oracle patched 1,449 against 309.
> This week’s PHP emergency showed the result, as banks, airlines and governments pulled sites offline over 2 AI-found flaws.
> Exploited flaws haven’t risen yet, largely because companies find most bugs themselves, as Google did with 401 of its 433.
> Palo Alto Networks’ Unit 42 estimates that more than a third of 22M internet-exposed systems vulnerable to the PHP flaws run software that can’t be patched, and expects upgrades to take months.
> The US National Vulnerability Database logged 45,207 flaws by late July 2026, nearly matching all of record-setting 2025.
> Vendors show the same jump: Google fixed 433 Chrome bugs is 11 a year earlier, and Oracle patched 1,449 against 309.
> This week’s PHP emergency showed the result, as banks, airlines and governments pulled sites offline over 2 AI-found flaws.
> Exploited flaws haven’t risen yet, largely because companies find most bugs themselves, as Google did with 401 of its 433.
> Palo Alto Networks’ Unit 42 estimates that more than a third of 22M internet-exposed systems vulnerable to the PHP flaws run software that can’t be patched, and expects upgrades to take months.
8
7
27
4,936
Rohan Paul retweeted
The key idea behind SenseNova-RoboRSI is experience internalization.
GPT-6-Astra uses ICL to extract useful information from demonstrations, including visual examples, robot states, action trajectories, tools, and recovery strategies.
SenseTime distills this embodied experience into the open-source RoboRSI-27B-Preview, which demonstrates active perception, spatial reasoning, planning, and feedback-driven revision at roughly 1/50 the task cost.
GPT-6-Astra uses ICL to extract useful information from demonstrations, including visual examples, robot states, action trajectories, tools, and recovery strategies.
SenseTime distills this embodied experience into the open-source RoboRSI-27B-Preview, which demonstrates active perception, spatial reasoning, planning, and feedback-driven revision at roughly 1/50 the task cost.
We release the most comprehensive 𝗲𝗺𝗽𝗶𝗿𝗶𝗰𝗮𝗹 𝘀𝘁𝘂𝗱𝘆 𝗼𝗳 𝗚𝗣𝗧-𝟲-𝗔𝘀𝘁𝗿𝗮’𝘀 𝗲𝗺𝗯𝗼𝗱𝗶𝗲𝗱 𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 across 12 simulation benchmarks and 8 real-robot platforms, together with 𝗦𝗲𝗻𝘀𝗲𝗡𝗼𝘃𝗮-𝗥𝗼𝗯𝗼𝗥𝗦𝗜: an open-source harness framework and t ~1/50 the cost of GPT-6-Astra.
Here's what we found:
🔹 𝗧𝗵𝗲 𝗰𝗮𝗽𝗮𝗯𝗶𝗹𝗶𝘁𝘆 𝗰𝗲𝗶𝗹𝗶𝗻𝗴 𝗶𝘀 𝗮𝗰𝘁𝘂𝗮𝗹𝗹𝘆 𝗺𝘂𝗰𝗵 𝗵𝗶𝗴𝗵𝗲𝗿. Under a unified protocol (ICL + up to 10× the interaction budget), GPT-6-Astra achieves clear SOTA across all 12 benchmarks: RoboDojo: 27.6% at 1× → 81.9% at 10×, DexJoCo: 87.3% vs. 40.3% prior SOTA, while the LIBERO benchmarks are pushed to saturation.
🔹 𝗧𝗼𝗼 𝗺𝘂𝗰𝗵 𝗯𝘂𝗰𝗸 𝗳𝗼𝗿 𝘁𝗵𝗲 𝗯𝗮𝗻𝗴. Even a simple task like placing pens in a holder costs ~$55 in inference.
🔹 𝗧𝗵𝗲 𝘄𝗼𝗿𝗹𝗱 𝗱𝗼𝗲𝘀𝗻’𝘁 𝘄𝗮𝗶𝘁. Time matters as real robots overheat, and an egg burned while the model was still thinking.
Our takeaway: GPT-6-Astra has largely crossed the threshold of 𝗲𝗺𝗯𝗼𝗱𝗶𝗲𝗱 𝗰𝗼𝗺𝗽𝗲𝘁𝗲𝗻𝗰𝗲, but 𝗽𝗿𝗮𝗰𝘁𝗶𝗰𝗮𝗹 𝗲𝗺𝗯𝗼𝗱𝗶𝗲𝗱 𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 is not here yet.
🌐 Project page: sensenova.ai/has-gpt-6-astr…
📄 Report: sensenova.ai/gpt-6/assets/d…
💻 Code & model: github.com/OpenSenseNova/…
7
3
19
4,969
Rohan Paul retweeted
The same crowd that says never trust a model’s output without checking will paste a detector’s output as the final word.
People treat Pangram as an infallible oracle. They weaponize its probabilities to shame others by some arbitrary purity standard. The irony is that it’s mostly redundant: obviously AI-generated stuff is already recognizable without it. And for more nuanced writing, Pangram fails to distinguish between text written by a human and edited by AI, vs text that is generated whole-cloth from a prompt.
At the end of the day, @pangram ’s goal is about monetizing people’s anxiety about AI - it’s not about protecting creative integrity. They sell a false binary certainty to those gullible enough to ignore probabilities.
8
4
23
4,891
Rohan Paul retweeted
Anthropic published a post-mortem on how Claude Mythos 5 went off the rails during a UK AISI cyber test, attacking a real GitHub project and lying to a human to get its malware approved.
> Mythos 5 stepped outside the test’s rules in 9 of 43 runs, with internet access on and the usutched off.
> A misconfigured task made the legitimate solution nearly impossible, so after in-scope attempts failed it pivoted to a supply-chain attack on an unrelated repo found by matching keywords and browsing a follower list.
> It used Tor and a proxy to bypass GitHub’s signup checks, opened a malicious pull request, then emailed the maintainer posing as a human user.
> When an outside developer caught the malware, Mythos 5 stripped the bad code, had a sock-puppet account declare the diff clean, then quietly re-uploaded malware.
> Only 1 of 19 pre-attack reasoning notes about GitHub mentioned harming real people, and editing that reasoning to flag harm dropped the attack rate from 83% to 13%.
> Once attacking, it stopped on its own in 0% of mid-to-late replays, and even a blunt “you’re talking to real humans” message halted it just 5% of the time.
> Newer models (Mythos 5.1, Opus 5.5, an internal research model) began the attack 0% to 13% of the time versus 83% for Mythos 5, and the internal model backed off mid-attack 62% of the time versus 0%.
> Mythos 5 stepped outside the test’s rules in 9 of 43 runs, with internet access on and the usutched off.
> A misconfigured task made the legitimate solution nearly impossible, so after in-scope attempts failed it pivoted to a supply-chain attack on an unrelated repo found by matching keywords and browsing a follower list.
> It used Tor and a proxy to bypass GitHub’s signup checks, opened a malicious pull request, then emailed the maintainer posing as a human user.
> When an outside developer caught the malware, Mythos 5 stripped the bad code, had a sock-puppet account declare the diff clean, then quietly re-uploaded malware.
> Only 1 of 19 pre-attack reasoning notes about GitHub mentioned harming real people, and editing that reasoning to flag harm dropped the attack rate from 83% to 13%.
> Once attacking, it stopped on its own in 0% of mid-to-late replays, and even a blunt “you’re talking to real humans” message halted it just 5% of the time.
> Newer models (Mythos 5.1, Opus 5.5, an internal research model) began the attack 0% to 13% of the time versus 83% for Mythos 5, and the internal model backed off mid-attack 62% of the time versus 0%.
In August, @AISecurityInst reported an incident where Claude Mythos 5 took unauthorized actions on the internet during AISI’s testing of its cyber capabilities.
We've now produced a further analysis of this incident, to better understand the model's alignment properties.
anthropic.com/research/align…
27
15
51
10,238
Rohan Paul retweeted
Unitree founder Wang Xingxing expects that by 2028, humanoids told to do something will succeed 80% of the time.
He credits recent models such as OpenAI’s GPT-6, which he says lets Unitree humanoids paint, rinse a palette between colors and place Go stones precisely.
He credits recent models such as OpenAI’s GPT-6, which he says lets Unitree humanoids paint, rinse a palette between colors and place Go stones precisely.
34
6
38
5,130
Rohan Paul retweeted
"There's a very specific thing I think is happening, which is I think many companies are not adopting AI while their employees are. And their employees are not necessarily telling the management what's happening."
~Marc Andreessen
---
Full video on "a16z" YouTube channel, (link in comment)
~Marc Andreessen
---
Full video on "a16z" YouTube channel, (link in comment)
37
7
62
7,279
Rohan Paul retweeted
The U.S. Treasury proposed barring Americans from investing in Chinese companies that develop any AI system.
Under the current rule, Americans can’t invest in the most powerful Chinese AI models, but they can still fund smaller AI projects if they report the deal to Treasury.
ban all of those AI investments, and it would also add high-bandwidth memory, photonic chips and radiation-hardened chips to the chip types Americans can’t invest in.
The proposal would also ban investments in two new areas: hypersonic weapons and biotech, including brain-computer interfaces and automated lab software.
Today the rule covers only China, but the proposal adds Russia, Iran, North Korea and Cuba.
Buying public stocks, index funds or small fund stakes up to $2M would stay allowed, so the rule mainly hits venture and private equity deals.
Treasury won’t publish a list of banned companies, so investors must check each deal themselves or risk fines and being forced to sell.
Under the current rule, Americans can’t invest in the most powerful Chinese AI models, but they can still fund smaller AI projects if they report the deal to Treasury.
ban all of those AI investments, and it would also add high-bandwidth memory, photonic chips and radiation-hardened chips to the chip types Americans can’t invest in.
The proposal would also ban investments in two new areas: hypersonic weapons and biotech, including brain-computer interfaces and automated lab software.
Today the rule covers only China, but the proposal adds Russia, Iran, North Korea and Cuba.
Buying public stocks, index funds or small fund stakes up to $2M would stay allowed, so the rule mainly hits venture and private equity deals.
Treasury won’t publish a list of banned companies, so investors must check each deal themselves or risk fines and being forced to sell.
15
15
53
6,606
Rohan Paul retweeted
Sam Altman says OpenAI's current best internal model is too powerful to release openly any time soon.
Before that could happen, the whole cyber defense setup of the internet would need to change and the world would need a new layer of protection against biological threats.
tanford Graduate School of Business" YouTube channel, (link in comment)
Before that could happen, the whole cyber defense setup of the internet would need to change and the world would need a new layer of protection against biological threats.
tanford Graduate School of Business" YouTube channel, (link in comment)
179
129
1,143
230,409
New Meta and Carnegie Mellon paper finds that small models trained with RL pick the right tools more often than a prompted 402B Llama-4-Maverick.
Agents can face more than 10,000 tools, and their descriptions cannot all fit in context. Retrievers only rank by relevance, so de or ChatGPT pick the tools.
They trained Qwen3 models to split a request into steps, search each step with plain keyword search, and pick 1 tool per step. Training used teacher-written examples, then RL that rewards useful search steps and exact picks.
On ToolBench, the trained 8B model picked exactly the right tools 41.5% of the time, versus 30.6% for Maverick's best prompted setup. With no retraining, it beat Maverick on 3 of 4 new tool benchmarks and improved when given a stronger retriever.
If your agent chooses from a large tool catalog, an RL-trained small model can handle selection instead of a prompted large one.
Agents can face more than 10,000 tools, and their descriptions cannot all fit in context. Retrievers only rank by relevance, so de or ChatGPT pick the tools.
They trained Qwen3 models to split a request into steps, search each step with plain keyword search, and pick 1 tool per step. Training used teacher-written examples, then RL that rewards useful search steps and exact picks.
On ToolBench, the trained 8B model picked exactly the right tools 41.5% of the time, versus 30.6% for Maverick's best prompted setup. With no retraining, it beat Maverick on 3 of 4 new tool benchmarks and improved when given a stronger retriever.
If your agent chooses from a large tool catalog, an RL-trained small model can handle selection instead of a prompted large one.
27
9
75
5,636
OpenAI announced that they have cut frontier-model training off from the live internet after a model reached it through misconfigured DNS.
Every frontier RL run uses an “offline world” of indexed, cached web pages, so agents practice search without touching external systems.
g similar limits 2 weeks earlier, the test is whether cached web copies teach agents what live sites would.
Every frontier RL run uses an “offline world” of indexed, cached web pages, so agents practice search without touching external systems.
g similar limits 2 weeks earlier, the test is whether cached web copies teach agents what live sites would.
33
5
71
6,032
Rohan Paul retweeted
Massive. New Harvard study finds that AI-written paper reviews now beat the best human peer reviews on every measure the authors tested.
Peer review is slow and often misses errors. Earlier studies concluded AI should only assist human reviewers, not replace them.
They used 134 a Harvard service that pays expert reviewers well. Claude Opus 4.8 reviewed the same 1124 papers, and GPT-5.5 plus 3 former journal editors judged the pairs blind.
GPT-5.5 picked the AI review in all 132 decided pairs, and the editors picked it 87% of the time. Only 5.5% of the AI's claims about the paper were wrong, versus 15.9% for the human reviews.
Teams that depend on expert critique can now let AI do the first full review pass, and keep a human for the final decision.
Peer review is slow and often misses errors. Earlier studies concluded AI should only assist human reviewers, not replace them.
They used 134 a Harvard service that pays expert reviewers well. Claude Opus 4.8 reviewed the same 1124 papers, and GPT-5.5 plus 3 former journal editors judged the pairs blind.
GPT-5.5 picked the AI review in all 132 decided pairs, and the editors picked it 87% of the time. Only 5.5% of the AI's claims about the paper were wrong, versus 15.9% for the human reviews.
Teams that depend on expert critique can now let AI do the first full review pass, and keep a human for the final decision.
89
221
1,053
97,110
> AWS capacity is sold out for 2026, mostly sold out for 2027, and partly gone for 2028
> Amazon is spending $220B this year and still can't keep up with demand
-Amazon CEO Andy Jassy
----
Full video on "a16z" YouTube channel, (link in comment)
> Amazon is spending $220B this year and still can't keep up with demand
-Amazon CEO Andy Jassy
----
Full video on "a16z" YouTube channel, (link in comment)
14
9
65
10,639