Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Claude 4

anthropic.com

Post reply on HN
Print view View original

111–120 of 1001 posts

Re: Claude 4

#111

Post by minimaxir »

hansmayer wrote: post #62
sigmoid10 wrote: post #18

Sooo... it can play Pokemon. Feels like they had to throw that in after Google IO yesterday. But the real question is now can it beat the game including the Elite Four and the Champion. That was pretty impressive for the new Gemini model.

Right, but on the other hand... how is it even useful? Let's say it can beat the game, so what? So it can (kind of) summarise or write my emails - which is something I neither want nor need, they produce mountains of sloppy code, which I would have to end up fixing, and finally they can play a game? Where is the killer app? The gaming approach was exactly the premise of the original AI efforts in the 1960s, that teac…

It's a fun benchmark, like simonw's pelican riding a bike. Sometimes fun is the best metric.

Re: Claude 4

#112

Post by SkyPuncher »

sali0 wrote: post #51

I've found myself having brand loyalty to Claude. I don't really trust any of the other models with coding, the only one I even let close to my work is Claude. And this is after trying most of them. Looking forward to trying 4.

Gemini is _very_ good at architecture level thinking and implementation.

I tend to find that I use Gemini for the first pass, then switch to Claude for the actual line-by-line details.

Claude is also far superior at writing specs than Gemini.

Re: Claude 4

#113

Post by throwaway314155 »

waynecochran wrote: post #43

My mind has been blown using ChatGPT's o4-mini-high for coding and research (it knowledge of computer vision and tools like OpenCV are fantastic). Is it worth trying out all the shiny new AI coding agents ... I need to get work done?

I would say yes. The jump in capability and reduction in hallucinations (at least code) to Claude 3.7 from ChatGPT (even o3) is immediately noticeable in my experience. Same goes for gemini which was even better in some ways until perhaps today.

Re: Claude 4

#114

Post by WillPostForFood »

MattSayar wrote: post #87
sali0 wrote: post #51

I've found myself having brand loyalty to Claude. I don't really trust any of the other models with coding, the only one I even let close to my work is Claude. And this is after trying most of them. Looking forward to trying 4.

Same. And I JUST tried their GitHub Action agentic thing yesterday (wrote about it here[0]), and it honestly didn't perform very well. I should try it again with Claude 4 and see if there are any differences. Should be an easy test [0] https://mattsayar.com/personalized-software-really-is-coming...

It would be pretty funny if your "not today. Maybe tomorrow" proposition actually did happen the next day.

Re: Claude 4

#115

Post by porridgeraisin »

throwaway314155 wrote: post #90
energy123 wrote: post #79

> Finally, we've introduced thinking summaries for Claude 4 models that use a smaller model to condense lengthy thought processes. This summarization is only needed about 5% of the time—most thought processes are short enough to display in full. This is not better for the user. No users want this. If you're doing this to prevent competitors training on your thought traces then fine. But if you really believe this is…

If _you_ really believe this is what all users want, _you_ should reconsider. Your describing a feature for power users. It should be a toggle but it's silly to say it didn't improve UX for people who don't want to read strange babbling chains of thought.

Here's an example non-power-user usecase for CoT:

Sometimes when I miss to specify a detail in my prompt and it's just a short task where I don't bother with long processes like "ask clarifying questions, make a plan and then follow it" etc etc, I see it talking about making that assumption in the CoT and I immediately cancel the request and edit the detail in.

Re: Claude 4

#116

Post by i_love_retros »

op00to wrote: post #30
i_love_retros wrote: post #23

[flagged]

Good point. We should only focus on intractable problems and put everything else on the back burner. We certainly don’t have the ability to help people and advance science and business.

How is this absurd gold rush advancing science? What important scientific discoveries have been made so far using LLMs? Chatbots? Coding tools?

The only business being helped is the few mega wealthy "AI" company owners.

This stream of announcements is a parody at this point.

Re: Claude 4

#117

Post by accrual »

Very impressive, congrats Anthropic/Claude team! I've been using Claude for personal project development and finally bought a subscription to Pro as well.

Re: Claude 4

#120

Post by vFunct »

SkyPuncher wrote: post #112
sali0 wrote: post #51

I've found myself having brand loyalty to Claude. I don't really trust any of the other models with coding, the only one I even let close to my work is Claude. And this is after trying most of them. Looking forward to trying 4.

Gemini is _very_ good at architecture level thinking and implementation. I tend to find that I use Gemini for the first pass, then switch to Claude for the actual line-by-line details. Claude is also far superior at writing specs than Gemini.

Yah Claude tends to output 1200+ line architectural specification documents while Gemini tends to output ~600 line. (I just had to write 100+ architectural spec documents for 100+ different apps)

Not sure why Claude is more thorough and complete than the other models, but it's my go-to model for large projects.

The OpenAI model outputs are always the smallest - 500 lines or so. Not very good at larger projects, but perfectly fine for small fixes.

Post reply on HN