dave

@ryanbrewer too early indeed

Sadly, too early is still wrong

Now imagine if you could do this on your local machine instead of a VM? @leonsilicon @wigglydtrain

ANTHROPIC DOWN CAN ONLY MEAN ONE THING

@ryanbrewer @leonsilicon i’ve been thinking this

We were just three years too early @wigglydtrain @leonsilicon

@gregpr07 do you like the nano texture screen?

There is something really fun about working at airports

@jschopplich @mattpocockuk ive been looking for and thinking about this for a long time

JSON is token‑expensive for LLMs – just like @mattpocockuk frequently mentions. Meet TOON, the Token‑Oriented Object Notation. 💸 40–60% fewer tokens than JSON 📐 readable & tokenizer-aware Wrap your JSON with `encode` to save half the token cost: github.com/johannschoppli…

@amasad So why’d you raise it?

Have *not* yet spent a single dollar from the money we raised 2.5 years ago!

@itsnotbernhard good on you for immediately turning to twitter

Just wrapped an interview with a candidate. Palo Alto born & raised, Palo Alto High school, Ivy League CS, internships at tier-1 startups & ai labs. On the call he literally melted down from the pressure. Parents, peers, expectations, he was completely overwhelmed. Palo Alto parents, please love your kids. Please put aside your own ego and just love them.

@benhylak you got enough apps there

no time to post. locked in from a mountain in alaska. talk to you all soon.

@jasonfreedman lmao

The Venture Series A industry will have a pre-Standard Capital and post-Standard Capital split. Exactly the same as the shift in seed funding that occurred when YC introduced the Summer Founders Program in 2006 and the SAFE in 2012. The market will shift over time to those venture firms with a Standard Capital-style application/fast decision process and those without. And those without will increasingly see that they didn't get a shot at a lot of great Series A opportunities like they used to. @daltonc and @paultoo are fabulous people and fabulous investors and will build a great firm and pull in top tier founders. They could have easily built a normal firm and raised this much based on the strength of their track record alone. But instead they chose to upend the whole model. It shows just how much they understand founders. No founders want to spend 3-9 months fundraising. No founders want to participate in fragile lead investor dynamics No founders want to negotiate bespoke Series A docs No founders want to pay for huge legal fees on both side of the transaction. No founders want to default sell 25% of the company No founders wants to default give up board seats. The reality for the moment is that founders will go first to the Series A firm that requires the absolute minimum amount of their efforts so that they can stay focused on their company instead. Currently Standard Capital is really the only firm that can offer that. Every Series A firm should be spending next week discussing whether they are going to adopt the Standard Capital approach. I hope many of them jump on board sooner rather than later.

@timkawakami do you think purdy throws that game winning touchdown away if he has greater confidence in jake moody?

I don't know what the 49ers will do this week with Jake Moody. I do know that Shanahan doesn't WANT to cut him--which was his emotion when I asked today. He's kept Moody through all other tough times. I suspect it will take a miss that costs them a game to push it over the line.

@ryanflorence @grok is this true

pov I find eslint in a typescript project I need to work in

@JohnPhamous very claire hughes johnson of you

a personal readme kinda like llms.txt but for humans

@nathan_covey because it doesn’t fucking matter

I'm sorry but if you're using Wix, Squarespace, or Wordpress for your website, how are you in YC??

@a_sarentorbic @thdxr knowing a ddd term doesnt automatically make u better lol. plenty of ppl like to say things like “bounded context” and “ubiquitous language” and still write garbage code

on this topic of structuring code for LLMs the truth is you're likely biasing towards technology you like and coming up with reasons why it's good for LLMs they're decent with most setups so you can find whatever proof you want if you want to really optimize try to shed all your biases - the root thing here is agents benefit from being able to consume a codebase granularly the reason codebases are less granular than they can be is because it's tedious - so most of our tech has traded off an amount of readability for simplifying authorship so the question is what patterns would you try if you weren't concerned about the pain of writing code and the answer are things that are generally unpopular - lots of files, separation of concerns, enterprise-y patterns

@gregisenberg never, ever give up equity in your company for a few hundred leads

Creator seed rounds are about to become a thing. Instead of giving ~15-25% of your company to VCs for ~$3M, you give that same chunk to a creator. For zero dollars. Sounds insane until you think about it. Standard seed round dilution is 15-20%. You need $3M to build and find customers, VCs want their ownership target, everyone knows the dance. But founders are starting to realize something, they don't need $3M to build anymore for software businesses. That's what doesn't keep them up at night Finding customers is what keeps them up. Enter the creator with 500k YouTube subscribers. Those subscribers are your exact target customer in your exact subniche. The creator already has their trust. One authentic mention from them is worth six months of paid ads. So you offer them 20% of your company. Same dilution as a seed round, but instead of cash you can't spend efficiently, you get customers you can monetize immediately. And a firehose of customer feedback (woo!). The YouTuber is stoked because sponsorships pay $30k once. Equity could be worth $30M. They'll talk about this product for years because they're an owner, not a rental. And they are constantly giving you feedback, more valuable than an investor because they are your target, and they want this product to be as valuable as possible to their followers. You can still raise from VCs later. But now you're raising a Series A at $30M with actual revenue, not a seed at $12M with hopes and dreams. The creator got you through the hardest part: proving people want what you're building. Or you just ignore the VC route. With customer acquisition solve, you become instantly profitable. Harder if this is a physical product but if you'e a software company (app, maketplace etc) this is the rule not the exception. Creator seed rounds. New trend. I think this will continue to be more common.

@gregisenberg lol

@SimGamingNetwrk @nut_history a’s are moving to vegas too lol

@nut_history The Rockies in the Pacific coast while the Checks notes Dodgers, Angels and padres Who are all actually on the Pacific Coast Are in the West. Smfh

Here are Jim Bowden proposed realignment Thoughts?

@thdxr if this is true about people making them, it’s probably true about the people using them what’re the main things people are doing wrong when using opencode?

looking through the source code of some of these coding agents i get the feeling the authors actually hate programming idk how well positioned they are to dictate how programming should be done

@thdxr it's because it wasn't able to find & replace. i think i'd probably expect it to update the search string instead of replacing my code

@thdxr hanged some of the styles on the skeleton, and decided to just change them back.