My apartment used to be on the other side of the river and I remember Tata Hall being one of the largest campus buildings in HBS (I believe it’s 2nd largest) Most of the MIT folks I would meet were also on Tata fellowships, apparently they fund a lot of graduate programs.

Can't believe a few hours of Fable set such high expectations for me, I feel 10x slower with the other models now : (

How does that even work? Models are non-deterministic, you can do your best to patch some patterns but unlikely to get rid of an ability entirely without severely impairing the quality of outputs. Defense in depth is the right approach here for now, maybe adding KYC helps in some way? I have a feeling the entire export control measure was taken only because that is the only tool the government had to make them shut it down immediately. A lot of businesses who treat software correctness, & security as theatrics will need to do a lot more work because that will be the only way they can avoid getting taken advantage of

They want humans that cost less than a Claude subscription! what a joke

Fable 5 is genuinely impressive, not just the benchmarks but on real hard engineering work It makes far fewer errors so you end up saving a lot of time overall as compared to GPT 5.5 It is also noticeably better at making good decisions when it encounters gaps in the implementation plan or when presented with new facts from an execution. If Anthropic can keep up with this pace, small engineering teams will be doing a lot more in-house than trivial apps. A motivated senior engineer can ship some serious tools with such models. It also validate's @mitchellh's take on the block economy, smaller well built primitives will become increasingly important and popular - databases, filesystems, VMs, memory engines etc. What this means for AI companies being built right now is hard to say, but it's definitely the best time to build the most ambitious thing you can think of.

We went from copy pasting code snippets to one shotting Replit within 2 years I wonder what 1 year from now will look like

If anything it’s an accumulated liability.

boomers think having a domestic help is a better quality of life

founder market fit

Tell me a more dramatic transition than this! I think it's super cool

This problem will naturally tend to go away as companies are grown from the start using AI. Then you don't need to extract any domain knowledge from people's heads; it will never have been in people's heads.

Solo-founded $2B company. Used by OpenAI, Anthropic, Replit. Solo Founders ep 13 is live with @grinich of @WorkOS. 07:49 The co-founder "blood bond" is a myth 10:55 There's only room for one Zuck at Meta 32:51 Pivots are the most traumatic thing you can do to a business 43:18 The best ideas always look like bad ideas 50:14 Sales is the ultimate expression of the company's value 53:57 The "founder pixie dust" no hire can replace 01:07:42 You only get one life. Why not swing as hard as you can?

founders before @supermemory

TigerBeetle came out of a real problem: - The design was sketched with a sharpie on a boardroom table, in about 30 mins, the day I first saw the problem (March 11, 2020). - But we then did 4 months of part time analysis to understand the problem more deeply (tracing all SQL queries and application logic of a central bank transaction processing system) and let the design simmer… - Finally, the first prototype of TB (to prove out the 1000x performance differential) was coded in a 5 hour sprint on a rainy Sunday afternoon in July 2020, sitting in front of the fire place, with a glass of red wine, listening to the Black Keys. That did 400K TPS, proving out the performance cost of the big ideas. From there it was a month to create AlphaBeetle in Zig. Then 90 days for BetaBeetle in Zig. Then TigerBeetle ever since…

I am going to piss off so many friends by saying this but if I could invest in one emerging manager sub $50M fund, it would be @Joshuabrowder. A few things you need to know about Josh: - He makes the founders he invests in live in his spare room at the Four Seasons until they raise their seed - He turned his $100K Thiel Fellowship grant into a $10M angel portfolio - He was one of the first cheques into Micro1, Yuzu and many more - When he found out his father had been taken by the Russians, he was playing poker… (legend!) I have never had founder references like the ones I got on Josh. I spoke to 12 founders. He averaged 9.2/10 across all 12. This is one of the best episodes we have done in a long time and my notes below: 1. Why I Believe Young Founders Make the Best Founders Young founders have no safety net and no option but to win. Corporate engineers often default to hiring big teams, while young founders stay focused on building the product. Their grit is much higher. Without that level of dedication, most people quit at the first real obstacle. 2. How I Test Founder Commitment Before Investing To filter out tourist founders, schedule a pitch meeting at 11:00 PM. Elite founders accept immediately. Mediocre ones push it out by weeks. During the interview, ask rapid-fire questions. If they claim a specific revenue number, have them pull up their live Stripe account on the spot. Look for tactical customer acquisition goals, not vague partnership promises. 3. Why I Make Founders Live With Me After Investing The best early investments come from deep day-one relationships. Living together creates a focused, one-person accelerator where founders get a three-week crash course and avoid years of mistakes. The rule is simple: co-founders share one room near the Four Seasons and cannot check out until they raise an institutional seed round. 4. Why Pre-Seed Companies Fail Startups usually fail for three reasons: they run out of money, they run out of hope, or the co-founders break up. Money problems usually come from weak pitching, which is why founders should drop the deck and show the product live. To maintain hope, ignore Silicon Valley vanity signals and focus on customer progress. To avoid team blowups, handle mechanics like vesting early. 5. What Founders Need to Know About Signing With a VC VCs will say almost anything to get you to sign on the spot. They reverse-engineer your desires and claim they know every customer you want to meet. Impressionable founders fall for it, but the promised intros often never happen. Never sign in the room. Take the night to think clearly. 6. My Biggest Lesson on Reserve Investing Holding back reserves for later rounds has a huge opportunity cost. The biggest value creation happens at pre-seed, so saving capital for a Series A follow-on can limit your upside. Deploying upfront into 20 to 30 pre-seed companies can produce far better long-term returns. Go all-in early. (links below)