The word is coming out…
Replying to @chrisbarber
(part 2) I asked people which neolabs under $10B have the highest talent density: Full list: frontierlist.com/neolabs 1 vote each: - Biography - Chai Discovery (@joshim5, @jackdent, @Mattmcpartlon1, @j_boitreaud) - CTRL-G (@GregoireGambatt, @pacovilletard) - Emulate (@jparkerholder, @MattMcGill_, @philipjohnball) - Engram (@dan_biderman, @EyubogluSabri, @RealJessyLin, @jxmnop, @scott_linderman, Chris Ré) - Factory (@matanSF, @EnoReyes) - Fulcrum - General Intuition (@PimDeWitte, @EloiAlonso1, @AdamJelley2, @micheli_vincent, Iggy Harmsen) - Isomorphic Labs (@demishassabis) - Olo Din (@mishayagudin) - Pantheon (@calixo888, @Jannikschg) - Sakana AI (@hardmaru, @YesThisIsLion, @renitomail) - Simile (@joon_s_pk, @msbernst, @percyliang) - Stealth - Stealth text-diffusion lab (ex-GDM) - TypeSafe (@CompleteSkeptic, @EGafni, @hackgoofer) - UniversalAGI (@aha_ml) - Wafer (@gpuemi, @gpusteve)
1
74
I asked a candidate this week to walk me through the rocket design process. His response: "you mean rocket science?" Rocket science will stop being rocket science very soon. It becomes a spec prompt. The next 5 years will be about AI for the physical world. Cars, rockets, drones will be designed, simulated, tested, and manufactured with AI. The market is so large that the LLM frenzy is just the beginning. And the business case is already simple today. Most people miscalculate the value as "fewer engineers," "fewer licenses," or "less compute": line items in enterprise budgets repurposed for a new vendor. That budget is easier to access, but cost reduction is the worst value prop unless you're selling a commodity. The reality: time to value is 95%+ of the value created. The classic study on this found that a product 6 months late to market earns ~33% less profit over five years. A product on time but 50% over its development budget loses only ~4%. So shrinking a design cycle from 12 months to 6 doesn't just save costs. It adds contracts. One additional program win generates billions, sometimes tens of billions, in revenue. The real question is how much of that upside you can capture. Even 10% gets you to $100M-1B per customer.
1
242
Ameer Haj-Ali retweeted
We've raised $700M at a $21B valuation from Jane Street, Kleiner Perkins, Sequoia, A16Z, Peter Thiel, BCV, and Blackstone. We're also excited to share that we've shipped our first rack to Jane Street.
306
453
6,977
7,823,270
Very well defined.
As a founder, your number one job is to eat shit. You're on the call with the upset customer. You need to fire people by telling them directly. You're listening to the engineer who wants to quit due to burnout. You have to read the 40-page vendor contract, which is a nightmare to go through. Every hard, uncomfortable thing lands on your desk. That's the whole job. You're here to absorb the bad stuff so everyone else can keep moving.
408
Ameer Haj-Ali retweeted
Today, we’re releasing a research preview of our real-time, local-first world model built for interactive, playable AI-worlds 60fps, locally run, all on consumer-grade hardware. Come take a look ⬇️
56
147
1,091
497,924
Somehow over the last few months, cursor bill went from $20/month to $600/day.
1
3
457
2026+ prediction: Physical AI will be bigger than @OpenAI !
6
427
Any reliable cloud provider that has on demand A100+ GPU availability?
2
396
Ameer Haj-Ali retweeted
.@RichardMCNgo on why birth rates will probably keep falling: - Birth rates are falling because marriage rates are falling. - AI could either distract us from relationships or help mediate them. - If the future economy is dominated by social + emotional labor, it may skew female - dating gets harder when women outpace men in status.
10
11
133
29,747
I might be right about a few things.
2
57
23,872
There is always an opportunity for positive outcomes 😂
Replying to @sama
no one of us gives a fuck who's right and who's wrong just please continue responding to each other so that we have more memes material
1
4
1,491
Jensen is an inspiring leader and role model. $5T 🚀 #NVDA #AI #Goat #Winning
3
455
Too much noise about RL it is impossible to find real signal
269
Have we tried GEPA + SFT? @LakshyAAAgrawal
Combining the benefits of RL and SFT with on-policy distillation, a promising approach for training small models for domain performance and continual learning.
3
1
17
3,035
Ameer Haj-Ali retweeted
Was running a terraform script in the middle of the AWS outage + thought I nuked our infra 🤣 Canva, Notion, and Asana down too
28
24
1,214
51,700
GEPA can augment even RL-tuned models! @IntelligenceArc team uses GEPA in their ATLAS framework, to augment an already powerful and RL-tuned teacher model, achieving +142% student performance improvement when guided by the improved teacher!
7
21
145
10,008
Exciting breakthrough!
Exciting progress on Middle East. Go @jaredkushner archive.ph/8Fxsm
355
Ameer Haj-Ali retweeted
Excited to share our new paper on AI-Driven Research for Systems. We show that AI can autonomously generate and verify novel solutions for classic systems performance problems, matching or exceeding human designs. A glimpse into how AI might transform not only systems, but the research process itself.
🚀 Excited to release our new paper: “Barbarians at the Gate: How AI is Upending Systems Research” We show how AI-Driven Research for Systems (ADRS) can rediscover or outperform human-designed algorithms across cloud scheduling, MoE expert load balancing, LLM-SQL optimization, transaction scheduling, and more — all within hours ⚡️ and under $20 💰. 🧵👇 Check it out!
5
34
176
31,912