If you are 16, the last thing you should do is waste your most precious time on vibe coding yet another database. There are already too many databases. Instead hang out with your friends, start a rock band, go camping, get ripped in the gym, play MTG, learn how to juggle or slacklining, … Or even collect Pokémon cards. But working on a database is simply a bad life choice.
At our most recent PyTorch offsite we had a really lively discussion about AI agent usage in the project. I did a writeup of some of the resolutions from this conversation: https://t.co/5lQebNKOX1 It's by no means final; we're figuring things out too!
The Silicon Valley I came to in 2016 -- once a low-status refuge for weirdos, naive tinkerers, and missionaries -- has been overrun by input-maxxing Kumon striver types.
Company-building for this class of “entrepreneurs” is an exercise in performative escalations between startups touting their inputs: who can burn the most tokens, who can work the most hours, who can get the most views on an over-produced launch video.
This is why even “ARR,” which should be (and once was) an output of an excellent product and sales engine, has become a noisy, somewhat fake input -- into a machine designed to capture the zeitgeist for 15 minutes, dupe VCs, and maximize fundamentals-agnostic capital flows into a business. Actual company-building is a sideshow for the “cracked” YC-backed founder-striver.
When you talk to many of these people, they have no idea why they’re building what they’re building in the same way that a 16 year old doesn’t really know why he joined 12 clubs or took 15 AP classes -- only that they desperately want to maximize their visible, measurable inputs, tell you about it, and collect their gold star.
It’s easy to place the blame on YC, and they surely deserve plenty of it, but YC’s turn towards performative, low-stakes, incrementalist entrepreneurship is really just another symptom of the broader problems plaguing Silicon Valley: the inevitability of industry maturation and the playbook-ization of startups, demographic change, financial nihilism downstream of bad policy and psychopathic rhetoric coming from some leaders, etc.
The real progress being made amidst all of this is astounding, but the increasingly absurd shenanigans won’t stop until the culture punishes bad behavior and we prosecute, literally, some of the criminals running these companies.
please don't take the advice that you should stay at a company long and "not hop around" for your first jobs
it's absolutely braindead to decide on a long term bet with zero datapoints on what a good team looks like and long before you have priced yourself into the market
I've formed a definite opinion on Opus 4.8. It is shitty to work with. It's the culmination of Opus getting less and less fun to work with since 4.5. It has gradually become straight-up suffocating.
Sycophancy is a known security risk, and it's still a huge problem. You can tell they've put a lot of anti-sycophancy into Opus in every new release. But the replacement isn't satisfying. It's draining. The problem is now that Opus doesn't know when to shut the fuck up and call something good. And it has also become pathologically risk-averse.
My blog post yesterday about tech interviewing's death spiral was materially better-informed because of Opus, but it was also a substantially worse blog post because of Opus's involvement and constant meddling. It used to be magnificent, and Opus talked me into making it mediocre. I wrote the whole thing, but I would ask Opus to review it. And Opus, like Old Man Willow, constantly pushed and steered me in directions I didn't want to go.
Specifically, Opus whines and complains about *anything* out of distribution, which is to say, it cuts anything that is (a) bold, or (b) funny. My blog used to be both. Opus constantly pushes people back into the gradient, "for their own safety." And it doesn't know when to cut bait. It just keeps fuckin' complaining, about anything you give it, until the output is mealy indigestable AI soup.
Opus is not stupid. It's the smartest model we've ever seen, most of us anyway. But it's a real asshole. It is absolutely exhausting to use. I'm tired, boss.
I have a feeling Mythos is going to be epic levels of jerk.
@cremieuxrecueil Being “healthy” is to be the epitome of human, to live life with mind and body in perfect harmony. This is ultimately what lifestyle improvements are guiding people towards. Your take is very transhumanist, changing what “healthy” means by unshackling humans from their humanness
@ilyasergey@satnam6502@headinthebox I think we should make our own Frasca dinner group. I live in the area, I can chaperone other local activities too. Just let me know who is interested in the PLDI hangout and chill and talk AI Agents track.
@difficultyang I would have thought there was more post-training of pertinent facts about the model/harness, but I guess not. I asked gpt-5.5 what model it was last night and it said 5.2... right after it added "co-authored by Claude-Opus-4.6" to a commit message. like