4 ms·

I don't know what I'm doing wrong but I've been pretty underwhelmed by o1 so far. I find its instruction following to be pretty good, but so far Claude is still

by causal 2y ago

I don't know what I'm doing wrong but I've been pretty underwhelmed by o1 so far. I find its instruction following to be pretty good, but so far Claude is still much better at taking coding tasks and just getting it right on first try.

crooked-v 2y ago
For me, Claude seems a lot better at understanding (so far as "understanding" goes with LLMs) subtext and matching tone, especially with anything creative. I can tell it, for example, "give me ideas for a D&D dungeon incorporating these elements: ..." and it will generally match the tone of theme of whatever it's given without needing much other prompting, while o1 will maintain the same bland design-by-committee style and often cloyingly G-rated tone to everything unless you get into very extensive prompting to make it do something different.
- causal 2y ago
  Claude definitely seems more "emotionally intelligent", but even just for 1-shot coding tasks I've been pretty bummed with 1o... like it will provide lots of output explaining its reasoning, and it all seems very sound, but then I run the code and find bugs that should have been easily avoided.
  - Alupis 2y ago
    My experience has been Claude is better at all coding/technology questions and exploratory learning. It's also night/day better at less common tech/languages than others (try asking ChatGPT questions about Svelte, for example). Claude is also vastly better with creative writing (adcopy) and better at avoiding sounding like a LLM wrote it. It's also vastly better at regular writing (helping you draft emails, etc). We were using OpenAI's Teams for a while. Tried Claude out for a few days - switched the entire company over and haven't looked back. OpenAI gets all the hype - but there are better products on the market today.
juped 2y ago
Claude is just better full stop these days, but you have to actually be attempting to get practical use out of various different models to know this.

by causal 2y ago

crooked-v 2y ago

For me, Claude seems a lot better at understanding (so far as "understanding" goes with LLMs) subtext and matching tone, especially with anything creative. I can tell it, for example, "give me ideas for a D&D dungeon incorporating these elements: ..." and it will generally match the tone of theme of whatever it's given without needing much other prompting, while o1 will maintain the same bland design-by-committee style and often cloyingly G-rated tone to everything unless you get into very extensive prompting to make it do something different.

causal 2y ago
Claude definitely seems more "emotionally intelligent", but even just for 1-shot coding tasks I've been pretty bummed with 1o... like it will provide lots of output explaining its reasoning, and it all seems very sound, but then I run the code and find bugs that should have been easily avoided.
- Alupis 2y ago
  My experience has been Claude is better at all coding/technology questions and exploratory learning. It's also night/day better at less common tech/languages than others (try asking ChatGPT questions about Svelte, for example). Claude is also vastly better with creative writing (adcopy) and better at avoiding sounding like a LLM wrote it. It's also vastly better at regular writing (helping you draft emails, etc). We were using OpenAI's Teams for a while. Tried Claude out for a few days - switched the entire company over and haven't looked back. OpenAI gets all the hype - but there are better products on the market today.

juped 2y ago

Claude is just better full stop these days, but you have to actually be attempting to get practical use out of various different models to know this.