Just tried Claude code. First impressions, it seems rather expensive. I prefer how Aider allows finer control over which files to add, or to use a sub-tree of a git repo. Also, It feels like the API calls when using Claude code are much faster then when using 3.7 on Aider. Giving bandwidth priority?
Claude 3.7 Sonnet and Claude Code
451–460 of 1001 posts
Re: Claude 3.7 Sonnet and Claude Code
#452Hi everyone! Boris from the Claude Code team here. @eschluntz, @catherinewu, @wolffiex, @bdr and I will be around for the next hour or so and we'll do our best to answer your questions about the product.
Serious question: What advice would you give to a Computer Science student in light of these tools?
Re: Claude 3.7 Sonnet and Claude Code
#453> strong improvements in coding and front-end web development
The best part
Re: Claude 3.7 Sonnet and Claude Code
#454Hi everyone! Boris from the Claude Code team here. @eschluntz, @catherinewu, @wolffiex, @bdr and I will be around for the next hour or so and we'll do our best to answer your questions about the product.
Your footnote 3 seems to imply that the low number for o1 and Grok3 is without parallelism, but I don't think it's publicly known whether they use internal parallelism? So perhaps the low number already uses parallelism, while the high number uses even more parallelism?
Also, curious if you have any intuition as to why the no-parallelism number for AIME with Claude (61.3%) is quite low (e.g., relative to R1 87.3% -- assuming it is an apples to apples comparison)?
Re: Claude 3.7 Sonnet and Claude Code
#455You can get your HN profile analyzed by it and it's pretty funny :) https://hn-wrapped.kadoa.com/ I'm using this to test the humor of new models.
> You've experienced so many startup failures that your LinkedIn profile should just read 'Professional Titanic Passenger: Always Picks the Wrong Ship'.
:'(
Re: Claude 3.7 Sonnet and Claude Code
#456So far only o1 pro was breathtaking for me few times. I wrote a kind of complex code for MCU which deals with FRAM and few buffers, juggling bytes around in a complex fashion. I was very not sure in this code, so I spent some time with AI chats asking them to review this code. 4o, o3-mini and claude were more or less useless. They spot basic stuff like this code might be problematic for multi-thread environment, thos…
Have you tried comparing with 3.7 via the API with a large thinking budget yet (32k-64k perhaps?), to bring it closer to the amount of tokens that o1-pro would use?
I think claude.ai’s web app in thinking mode is likely defaulting to a much much smaller thinking budget than that.
Re: Claude 3.7 Sonnet and Claude Code
#457It's pretty fascinating to refresh the usage page on the API site while working [0]. After initialization it was up to 500k tokens ($1.50). After a few questions and a small edit, I'm up to over a million tokens (>$3.00). Not sure if the amount of code navigation and typing saved will justify the expense yet. It'll take a bit more experimentation. In any case, the default API buy of $5 seems woefully low to explore t…
Update: Code tokens appear to be cheaper than 3.7 tokens, looks like it is around $0.75/million tokens for code, rather than the $3/million that the articles specifies for Claude 3.7
Re: Claude 3.7 Sonnet and Claude Code
#458You can get your HN profile analyzed by it and it's pretty funny :) https://hn-wrapped.kadoa.com/ I'm using this to test the humor of new models.
> You've spent more time justifying your Apple Vision Pro purchase than actually using it for anything productive, but hey, at least you can watch movies on 'the best screen' while pretending it's a 'dev kit'.
blasted
Re: Claude 3.7 Sonnet and Claude Code
#459Earlier quoted context omitted.
This test has always been so stupid since models work at the token level. Claude 3.5 already 5xs your frontend dev speed but people still say "hurr durr it can't count strawberry" as if that's a useful problem
This test isn't stupid. If it can't count the number of letters in a text, can you rely on it with more important calculations?
You can rely on it for anything that you can validate quickly. And it turns out, there are a lot of problems which are trivial to validate the solution to, but difficult to build the solution.
Re: Claude 3.7 Sonnet and Claude Code
#460So far only o1 pro was breathtaking for me few times. I wrote a kind of complex code for MCU which deals with FRAM and few buffers, juggling bytes around in a complex fashion. I was very not sure in this code, so I spent some time with AI chats asking them to review this code. 4o, o3-mini and claude were more or less useless. They spot basic stuff like this code might be problematic for multi-thread environment, thos…
Claude 3.5 Sonnet is great, but on a few occasions I've gone round in circles on a bug. I gave it to o1 pro and it fixed it in one shot.
More generally, I tend to give o1 pro as much of my codebase as possible (it can take around 100k tokens) and then ask it for small chunks of work which I then pass to Sonnet inside Cursor.
Very excited to see what o3 pro can do.