camelmel · HackerTrans

el mes pasadodiscuss

Yes, I can churn out a lot more stuff as can most of my peers. Experiments etc are all way faster to run with coding agents. But I think the overall creativity and originality is a lot lower. I think this is what many people are facing, if you don't use LLMs your short term productivity is worse.

camelmel

el mes pasadodiscuss

I didn't say I'm immune to those effects, I'm including myself in this as well. (also, I'm not older than my colleagues).

Most people definitely can't meditate for 30 minutes, so if you can do this, it's very impressive. Regardless, being able to think about poorly-defined problems and build completely new mental models from nothing is genuinely a really hard and uncomfortable task. If you don't use the skill you'll lose it.

camelmel

el mes pasadodiscuss

I have some sympathy for these kids. If LLMs were around when I was a student, I would've also used them to "speed up" my homework assignments then proceed to fail all my tests.

Now I work mostly with PhDs who were at the top of every academic environment they've ever been in. And yet I can see their thinking skills rapidly declining as well; many of them can no longer brainstorm, code, think deeply, or write without an LLM present doing 90% of the work. Many of them can no longer sit quietly for even 30 minutes just thinking on their own, which is a required skill for producing original thought.

For adults the cognitive decline won't be as measurable since there's no exams, and overall output volume will still be fine due to LLM help. But I do believe it's already happening absolutely everywhere around us. Honestly, I wanted to be in denial about it before but it's too obvious to ignore now.

camelmel

el mes pasadodiscuss

LLM written article. It's also not accurate; the fact that language models have human-interpretable representations and neurons has been known since BERT.

Circuits research also does not come from Anthropic. Mech interp is a huge field in academia and most of the core circuit analysis papers were from OpenAI/GDM/academia. However, Anthropic tends to produce a lot of blog posts where they draw poorly supported but hype-able analogies between LLMs and biological intelligence. It's wild.

For a better understanding of mech interp and circuits, including what we actually do know about LLM internals, I would recommend reading this paper: https://arxiv.org/pdf/2501.16496

camelmel

el mes pasadodiscuss

Huh, according to that model card this is a 137B total parameter model.

Performance doesn't seem that good:

- MAI-Code-1-Flash (137B-A5B) = 51% on SWE-bench pro

- Qwen3.6-35B-A3B = 49.5% on SWE-bench pro (https://huggingface.co/Qwen/Qwen3.6-35B-A3B)

They benchmark against Claude Haiku but Haiku is not good, it's worse than tiny open models you can run locally or via API at 10% the cost.

camelmel

hace 2 mesesdiscuss

Is this training data even valuable? Usually AI data annotators get paid to write LLM responses, but here all they'd be getting is a bunch of user queries.

comments