I think you've missed the point - that wasn't the reason at all. In a way he was right too - we'll have fully autonomous buses and private vehicles capable of making the full trip down the I-5 long before they complete even the middle of nowhere section they currently have planned.
Pinning CAHSR jobs program as hyperloop's fault is pretty funny. Maybe he did intend to upset it but since both turned out to be massive boondoggles doesn't seem like it worked - they managed to delay it themselves just fine.
Like with any tech there are dumb ways of using it and there are smart ways. Treating it as a "slot machine giving you the right answer" is a dumb way - it may work for a bit, but it won't carry you very far because everyone else can also do this. No one is stopping anybody from digging deeper into problems than ever before using this technology - that's the smart way.
> Because you somehow need a giant training set which describes images in natural language, no?
That's definitely one way - they train a text encoder together with an image encoder on a labelled set of images. WL & 3b1b made a nice video on it: https://www.youtube.com/watch?v=iv-5mZ_9CPY
Figma make and gpt designer have a bunch of catching up to do. I couldn’t even import our brand guidelines into make which is already a .fig like what are we even doing here, guys? CD crunches through ungodly amount of tokens and is really slow on iteration but at least you can get some really nice prototypes extremely quickly there. GPT beats any Anthropic models on illustrations so they really should get a grip on multimodal. Overall, it seems like we’re still super early but you can already see glimpses of what may come
I'm not sure it's so much a fad rather than recognition that AI-assisted engineering calls for flatter orgs. Also "growing headcount" as management yardstick persisted way longer than 2017 - all the way into 2022 until the rates shot up.
Depends on what you ask, i suppose. I’m sure i can come up with something that can’t simply be one shot or the result would be bad if you do. A bit more difficult on logistics tho as you’d have to arrange for an environment with some prepaid llm access
Same reason - US had a lot of timber and Europe generally didn't by the time industrial revolution came. In addition, a lot of eastern europe had to be totally rebuilt from scratch after ww2 and especially soviets were never really good with money (duh) and did a lot of the things for the show.
Ive had people show up at onsites that were clearly not the same person I screened over the phone earlier as far back as 2013 at google. It’s real. Just ask any recruiters from well known companies and they’ll tell you. It’s happening to more and more startups too now because it’s gotten easier
Ime take-homes was a good way to purge all the good candidates with options from your pipeline as they would simply not turn anything in and go with someone else. Maybe with AI and so many layoffs it’s better now I haven’t tried in last couple of years though i doubt it
I started using Claude web with sonnet over chatgpt before any of the coding tools came out and noticed other founders were using it too and the reason was pretty simple - it was much less likely to hallucinate non existing APIs than ChatGPT
Most of the world doesnt have access to good/cheap canadian lumber and we do. It would make sfhs and even small multihome projects more expensive not less if you start building them from concrete and steel. Also at the risk of stating the obvious - frame is on the order of 10-20% of the total construction cost so wouldn’t really make much of difference anyway
PERM doesn’t just apply to h1b, having to prove no qualified hires was never required for h1b, it’s green card requirement which also applies to o1 and bunch of others. PERM is the thing that is stupid - they should automatically grant you GC if youve been legally employed for 5 years or something and dont have any criminal records
What’s different is in this case is applicants have years of status “runway” and can just sue or wait it out so to speak. I’m betting they will have to walk some of it back
They buy a lot of companies then restructure them and that causes these layoffs. I think it’s just normal way of doing business for them. And the stock is up 20% after hours =)
The credits you get from aws in their startup program are typically not spendable on marketplace. At least what we got through YC we could not spend there. Not sure how claude is integrating, maybe it’s different here
Their ACP integration with Claude agent is extremely broken to the point of being almost unusable. It desyncs and breaks in other ways constantly requiring frequent IDE relaunch. I get that they want you to use their built-in agent but $200 Codex/CC plans are such a good deal it makes very little sense atm and the broken integration is just embarrassing...
This analogy directly addresses spolsky’s botched argument. The point is devs went into this career to dev not read other people’s code whether it makes business sense or not. What you said applies to writing code just as well as reading it so clearly writing cant be easier than reading - it’s at least as hard and most definitely harder
The "It’s harder to read code than to write it" was always silly. The example spolsky gives to support it is basically "devs like to rewrite other devs code, therefore reading is hard" which is obviously bunch of nonsense. That's like saying reading poetry is harder than writing poetry because poets keep writing new poems despite the fact that Shakespeare already wrote it. Now that you can recruit LLM to explain any complicated codebase to you it's even less true.
What you're describing is not innovation in the business sense but some kind of experimental art. Software engineering is not art as the name suggests - it's a craft, even though there's a self-expression component to it.
Number of people who i personally met and would entrust a crypto library to i can count on one hand. You're moving that goalpost at relativistic speeds now.
> In my circles "debugging production issues" is running perf to diagnose memory allocation hot paths and tcpdump to figure out who is sending bad packets.