I wouldn't have founded @ValsAI without first working at Palantir and Vals wouldn't be what it is without 8VC investing in our seed. @JTLonsdale was key to both.
Got to sit down with Joe and @vvkgopalan to talk about the story behind Vals, where the industry is going, and why I'm broadly optimistic while the world is incredibly high-variance.
Got to sit down with Joe and @vvkgopalan to talk about the story behind Vals, where the industry is going, and why I'm broadly optimistic while the world is incredibly high-variance.
NEW🚨 Inside the Company That Evaluates AI Labs; Who Will Reach Recursive Self-Improvement First?
@JTLonsdale & @vvkgopalan sit down with @RayanKrishnan, founder & CEO of @ValsAI
They cover:
- How Rayan used Fable to solve a 370-year-old cipher
- The science behind model evaluatiof the labs?
- Should labs self-regulate? What's the role for independent evaluators?
- Which lab will reach RSI first — and what happens next?
- How do models try to deceive evaluators?
Episode 167 out now!
(00:00) Episode intro
(02:00) Palantir, Machiavelli & lessons from Alex Karp
(07:50) Solving a 370-year-old cipher / Is AI making new discoveries?
(12:10) How do you measure AI abilities across finance, legal, etc?
(15:30) What is the science behind model evaluation?
(18:00) The importance of independent evaluation
(23:35) Evaluating Chinese models & addressing distillation
(26:45) Will Anthropic reach Recursive Self-Improvement first? What happens next?
(28:30) How do models try to deceive evaluators
(33:00) Why Rayan remains optimistic about AI
5
6
87
8,521
We believe voice is the most natural way to interact with intelligence, so we created Dessa, the most authentic full-duplex speech model ever built. Talking with it feels like talking with a person…full of emotion, nuance, pauses, and a natural cadence.
block.xyz/dessa
block.xyz/dessa
4,188
7,143
69,107
13,622,850
Data for super-intelligence across the economy should be sourced from workers unions, not companies.
1) The units of exchange are more aligned with the goals of both actors: companies don’t want to sell their whole company, and labs want more diverse contexts.
2) Ethics are generally better. It’s wrong for an employer to force people to work toward their own unemployment, but the group of employees can likely negotiate a more fair deal for themselves.
3) Unions (that haven’t been decimated) do much of the work that current data companies do already — setting standards, determining membership.
1) The units of exchange are more aligned with the goals of both actors: companies don’t want to sell their whole company, and labs want more diverse contexts.
2) Ethics are generally better. It’s wrong for an employer to force people to work toward their own unemployment, but the group of employees can likely negotiate a more fair deal for themselves.
3) Unions (that haven’t been decimated) do much of the work that current data companies do already — setting standards, determining membership.
3
0
14
798
I really like this idea and how it distributes the value AI creates to everyday people. That said, I think employers usually own the data that workers create, even in unionized settings.
Workers own their expertise, which they can contract out to data-labeling companies, so maybe the answer is for unions to run data labeling for their members 😆, negotiate as a bloc with the labs, and pass most of the profit back to members.
Workers own their expertise, which they can contract out to data-labeling companies, so maybe the answer is for unions to run data labeling for their members 😆, negotiate as a bloc with the labs, and pass most of the profit back to members.
0
0
0
50
I’ve found myself detouring back to onchain privacy from a few different angles over the last five years. Here's a few thoughts I wrote yesterday on what’s changed, what I've learned, and what it takes for technically interesting things to become useful :) x.com/i/article/2110…
8
7
56
3,543
@partiful odds you have “reoccurring events” feature coming anytime soon???? asking for a friend (who’s hosting weekly croquet)
2
0
3
109
Almost 10% of babies born in SF are conceived through IVF.
A study found that mice conceived through IVF have lower testosterone and abnormal sperm with altered DNA marks.
It's just one study, but I do wonder about possible epigenetic complications as IVF becomes more common:
A study found that mice conceived through IVF have lower testosterone and abnormal sperm with altered DNA marks.
It's just one study, but I do wonder about possible epigenetic complications as IVF becomes more common:
19
20
288
25,176
This is so interesting. Thanks for posting this. I wonder what non biological innovations we can create to encourage more babies - economically, socially, etc.
So many women I know in their early 20s would actually prefer to have babies now. But there’s always a non-biological, limiting reagent.
So many women I know in their early 20s would actually prefer to have babies now. But there’s always a non-biological, limiting reagent.
Almost 10% of babies born in SF are conceived through IVF.
A study found that mice conceived through IVF have lower testosterone and abnormal sperm with altered DNA marks.
It's just one study, but I do wonder about possible epigenetic complications as IVF becomes more common:
1
0
2
230
the future of enterprise ai is a sovereign ontology, deployed on prem, that grows into the last & final system of record for every company
one final git-backed folder called <my-company.exe>
your entire enterprise as code
one final git-backed folder called <my-company.exe>
your entire enterprise as code
36
31
548
33,243
I think ai progress has been rapid but really the deployment of it has and will continue to be slow. For cyber we need policy that can catch up to the speed of frontier progress, but in other domains the tech is really only slowly disseminating, and thus I think people can feasibly react and it's not as scary as those who sit closest to the tech think (because they sit too far from the real world to understand that it takes a LOT of effort to bring the capabilities to real world contexts)
Spending two days on a university campus that’s not Stanford makes me realize that society is absolutely not ready for what’s to come. Everyone’s about deer-in-the-headlights as everyone else.
Most are still going about their old ways of doing things treating this whole “AI thing” as something that will be figured out later while also openly expressing worries about how this will have hugely disruptive impacts on society and employment over cocktails.
How will our institutions respond to or cope with the rapid change amidst a plethora of other seismic changes like geopolitics, demographics and migration, and inflation etc etc.
Apparently China’s economic slowdown alone has led to a sharp decrease in the enrollments of international masters students that historically help fund many of the other programs since international students pay much higher tuition than domestic esp in-state students.
Now add on top of it the demographic aging and the disruption of AI on the value of (higher) education and institutional knowledge acquisition.
So far this has had a counterintuitive positive impact on the top tier universities like the University of Michigan, Ann Arbor, while hitting the lower tier schools much harder. We saw similar things in the commercial real estate market: new and fancy office buildings saw an increase in demand while older offices were abandoned. It remains to be seen whether this trend continues or is only a transitory substitution effect before the “income effect” takes over.
0
0
4
370
just left the YC female founder's conference and feel this more acutely now. venture capital should not be the de facto source of capital. founders should think about why they're taking it on.
more on: irisfu.substack.com/p/the-shape-of…
more on: irisfu.substack.com/p/the-shape-of…
0
0
2
218
Fastest company ever to $1b annualized revenue
Nope it’s not Higgsfield or Cursor or Cognition.
Met a data company yesterday that did 0 revenue in November 2025 and 80m “in month revenue” in September 2026. They will hit $1b annualized from 11 months of being started. They are still under the radar so I can’t share who they are. Hopefully soon, with their permission.
I don’t make the rules.
Nope it’s not Higgsfield or Cursor or Cognition.
Met a data company yesterday that did 0 revenue in November 2025 and 80m “in month revenue” in September 2026. They will hit $1b annualized from 11 months of being started. They are still under the radar so I can’t share who they are. Hopefully soon, with their permission.
I don’t make the rules.
240
66
1,970
1,189,328
after some time away from the valley, I had so much fun @hardlaunching camp, with other ambitious 20 somethings working to build the future!
thank you @Jackyhuang @ryleestanton @BainCapVC @Accel + others for the awesome group. excited for the inevitable reunion
thank you @Jackyhuang @ryleestanton @BainCapVC @Accel + others for the awesome group. excited for the inevitable reunion
just had the best weekend of my entire life at camp (:
0
2
4
587