Rajesh Jayaraman

@WillManidis I dunno man. You could have called the performative grindset crowd for what they are. Losers. Well written, though

@AjitPai Missed the chance to make a bouquet reference!

Note for less-cultured Europeans: you can only call it Kirkland Cabernet Sauvignon if it comes from the Costco region of the United States.

I now have @claudeai code down to $20 plan mostly for continuity (was at $200 last month, $100 this month). @OpenAI codex is now on the $100 plan. Might move to $200 as needed. Codex is just better and nicer to use.

@MatthewSitman He actually wears a coat and tie and walks to his office every morning and comes back in the evening. So no, he does not take afternoon naps. (Working by Caro)

Gang, I’m just having a little fun here, I’m sure Caro takes afternoons off sometimes, and I’m a big fan of Jeff’s work

You think Caro takes afternoons off?

@jhong So you are saying, that eli lilly giveth what anthropic takes away…

Would you rather be a smart engineer or a dumb but very hot person? (And yes one can certainly be both but outside an infinitesimal group, the ultra wealthy, let's assume the two traits are mostly orthogonal) I used to ask smart people I knew this question and they would often say smart. But I came to realize from meeting lots of hot people that actually the smart people didn't know what they didn't know. To be hot is like living in a world where you are lucky all the time. People are a lot nicer to you and go out of their way to help you. People sometimes literally just give things to you. (Symmetrically, many hot people do not realize that life is harder for the people who are not hot, their model of the world can be equally oblivious to the other side) Btw being can cause problems too, in the way that being too smart can also cause problems for people. My point was just that the answer is not as clear as people think. Being dumb and hot might arguably be better. And that's a good thing, because in a scenario where AIs come in and take over everything, their population will grow to enormous proportions and humans will become scarce entities... and AIs trained on human data might find that scarcity attractively. That means in the future, we humans will all be the dumb but hot ones. So... would you rather be a smart engineer or a dumb but very hot person?

@regulatorynerd Happy to. Many spv mouths to feed

@rajeshj Can you make it 50% so it mirrors the triple layer SPV fees I’m paying on my Anthropic gains?

I have a small allocation in Anthropic via a triple-layer SPV. I am looking to place a similarly small investment in Figma via a triple-layer SPV so I am perfectly hedged regardless of which company wins the design space. Please DM me if you are running the third layer of such an SPV so we can arrange my investment.

@rohitdotmittal With that rev, gm (assuming those revs and gm are correct and not magic math) and staffing they should be at or close to breakeven. If they have fixed cost overhang they should cut it and become a proper going concern business?

Talked to a founder this week: Raised: $10.4m across seed and Series A Last valuation: $42m post Current ARR: $2.1m Growth last 12 months: 4% Gross margin: 71% Team: 9 Runway: 12 months of cash Years since Series A: 3 This company could be a real business, but it'll never return the fund. The founder keeps grinding, thinking that his investors care. But his investors have already moved on to companies that are working. Some founders also like the "founder lifestyle." They stay stuck with low salaries because they want to be called founders, and not employees. Nikesh Arora shared what Masayoshi Son told him: "You’re spending too much time on the mistake. Don't bother. Your time is better spent with the winners." The investors keep not writing the company off because it's a call option for them. If it works, great; if not, they have others. But the team keeps showing up and grinding to return the invested capital. Any sale will not clear the prefer stack, so they don't sell. They can't shut down because it's a running business with customers and a team. They can't raise - 4% growth doesn't justify it. They can't quit - they've been underpaid for five years and have no savings. They are stuck - metaphorically and structurally. Helium Ventures is a standard exit path for founders stuck on the wrong cap table. The goal isn't to be the only option on the table. It's to be one of the good options, so founders can actually choose.

Codex computer use is incredible! Great work @OpenAI

@rayansadri 1.2B aum wealth firm doesnt make 1.2b revs or valuation

Dude’s company got merged for $1.2B, then he casually went back to doing side quests at the MIT lab. Next level badass

@paulg Same goes for misery too. *waves hand at everything*

Taught 14 yo (he's now 14) about the concept of a conservation law, and how doing good things for people seems to violate one in the sense that it makes them happier but makes you happier too.

@tlbtlbtlb Trevor - we just released flexskills.ai. That maybe able to make this a bit less painful. Would love for you to take a spin and see if you like it.

Once again, it's time to download 25 K-1s from 5 clunky investor websites and upload them to my tax guy's portal, hoping I don't miss any or risk huge penalties.

@ashleymayer @AveryAlchek Harlem

@AveryAlchek I'm interested in other people's opinions so didn't want to bias...but short list is good restaurants/grocery shopping, pet friendly, easy commute to lower manhattan.

NYC friends! We need to move apartments in the next 2-3 months (ugh), and I'm curious what you think are the most awesome + under-hyped neighborhoods? We currently live in Fort Greene and love it, but supply is low (and prices are crazy) so we want to cast a wider net.

@braveben Been very happy with delonghi dinamica fully automatic.

What’s the best espresso setup for home/small office use? Ideally $5k or less but open to ideas No complicated maintenance or excessive cleaning A beautiful object

@andrewparker High discount rate

Dating apps are such funny businesses. When they work, the value to the consumer is huge. But you don't really know until like 5 years later, and long after you deleted the app. Insane levels of consumer surplus over pricing.

Shouldn’t this be “Mici”?

@CJoeBlack i.giphy.com/l2JhCYVlbiCCxC…

It’s always “omg the bill is so high we have to do something about those guys and their industry.” It’s never “wow thanks so much for designing the option pool. I was able to raise a $30M series A with the available cap table room from people who wanted to wire money right away.”

@OGTaylorB @SketchesbyBoze @wmarybeard Good sentence: A child alone weeping at the slave market is a sermon on slavery that requires moral degeneracy, not literacy to misunderstand.

Thinking people in the past were too primitive to understand things like murder or slavery being wrong is such an arrogant thing to believe. As if ethics are like technology and build upon itself. You need Bronze Age ethics to get Iron Age ones. Nonsense. A child alone weeping at the slave market is a sermon on slavery that requires moral degeneracy, not literacy to misunderstand.

As @wmarybeard notes in “SPQR,” the slaughter of a million Gauls sparked a massive public outcry. Cato the Younger suggested Caesar be handed over to the Gauls for trial. Pliny the Elder declared him guilty of “a crime against humanity.” Please read even one book.

@M_C_Klein Depending in how complex the extraction is, you may want to give gemini a schema for structured extraction. Lmk if you need help!

@rajeshj will give it a try!

AI anecdote: tried to get Gemini to create an excel spreadsheet from a scanned PDF of old statistical tables. It wouldn't even try. Asked Claude and it made something where the formatting was nice and a lot of the numbers were correct, but it also got a bunch of things wrong.

@M_C_Klein Claude is bad at data extraction from pdfs - gemini is good. Gemini is bad at excel work, claude is good 🤷‍♂️

@rajeshj basic web claude (part of this was just trying to see if it could work)