Hacker News new | past | comments | ask | show | jobs | submit | MaxLeiter's comments login

We've been working on solving a lot of these issues with v0.dev (disclaimer: shadcn and I work on it). We do a lot of pre and post-processing to ensure LLMs output valid shadcn code.

We're also talking to the cursor/windsurf/zed folks on how we can improve Next.js and shadcn in the editors (maybe something like llms.txt?)


Thanks for all the work you do! v0 is magical. I absolutely love the feature where I can add a chunky component that v0 made to my repo with npx

We don’t know o1 thinks in English. What we see is a summary of the thinking process.

My guess is it “thinks” in non-eligible tokens


The o1 release blog post contains 8 full examples of o1 chains of thought (not the summarized versions visible to users). They're English.

https://openai.com/index/learning-to-reason-with-llms/#chain...

I have seen the summaries dip into completely random languages like Thai, so it might switch between languages occasionally.


Did you mean to write "non-legible" ?


Were you using claude’s computer mode? It can do this


No, I used the regular Claude which can also (somewhat) do this and uses the same image processing backend as "computer use" as far as I know (source: Antropic CEO interview with Lex Friedman recently).

Computer use is also not very good at it (often mis-clicking for example).

I'm guessing this will work flawlessly within 6 months to a year or so, but it doesn't seem ready yet.


Hey there, it's working fine for me on Firefox. Can you share more about your browser version/other details?


This confusion came from me not thinking too long before writing. My work makes use of feature flags so we incrementally merge features all the time, internally (and can do it quickly). But its still “shipped.”


Do I care? No

Does my manager? No, but what if they leave and I have a new manager right before performance reviews? I’d hope they don’t care about it, but I assume they’ll see it. May as well make it look nice.

Does the random recruiter looking at your resume care?

At some point in your career you might be able to afford to not care about this. Most people don’t have that luxury.


I wouldn't want to work for a company where the manager cares about green pixels tbh. But maybe I have that luxury you are talking about, I don't know


Me neither. But companies (and people) aren’t perfect. Your situation can change any day.


Pretty sure the first thing that random recruiter is going to do is go "wow, this guy worked at Vercel. Can we even afford him?"

I get that the market is pretty tight at the moment but strong CV positions have always trumped proverbial internet points.


If people invested as much time improving their abilities as they did gaming the system, they might find the problem solves itself.


Right - you fail at leetcode interviews despite your high abilities, and leave programming behind in anger to become a gardener.


I left off a key point; your code doesn’t need to ship to production, I just think you should do _something_. Thanks for sharing your thoughts.

> I can personally say that the satisfaction is far bigger the longer you've persevered.

It’s a great feeling to finish something. I find I’m more likely to finish it well if I break it up versus ship it all at once (and my teammates thank me)


I was about to say; if we replace 'ship' with 'commit' I'm in full agreement.

If you're a full time developer I don't know why you wouldn't.

Commits are communication. If we're making them often and correctly, there's that much less fluffing around with reports and status updates and meetings that we need to do.

The merits of CI/CD notwithstanding I definitely do NOT want my reports feeling like they need to deploy every day. That will lead to rushed work and errors on production.

But commit, why not? I don't care if it's in a private branch, behind a feature flag, or even some notes/pseudocode in a comment, commit it. Write an actual decent commit message while you're at it. That way as your manager, I can call up your commit history before our 1:1, and by the time the meeting starts it's already 80% finished because I was able to update myself on what you've been doing.


Managers watch commit histories? Haven't ever heard of something like that.


I don't know about routine skimming of history, but promo committees in big tech look at both commits and code reviews (the comments you left on others' commits).


I work at a FAANG like company, and here it's just annual performance reviews, that include bunch of achievements and then politics. No way anyone is going to check anyone's commit history. There are huge calibration meetings where managers have to stand up for their engineers on who gets what kind of performance rating and who gets promoted, and it's just verbal debates. There's just no way anyone can bring up someone's commit history and have enough time to explain what is going on overall.

And managers have little clue about the actual work that is going on. Definitely not enough clue to understand what commits are about, even if they have had a technical background overall.


  ; At this time, for DOS 4.00, we only have maximum 11 bytes left
  ; for translation.!!!!!!!!!!!!!!!!!!!!!
https://github.com/microsoft/MS-DOS/blob/main/v4.0/src/BOOT/...


Very cool project. I noticed in the dependencies section its using its own JS interpreter: https://mujs.com/


For an iOS attempt, I've often gone back to this wonderful StackOverflow post: https://stackoverflow.com/a/10289913


Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: