Viewing profile — Amekedl

You got to know how to use the model+harness+prompt to achieve the results you want, but honestly for many projects and questions all the models already pump out their same best ve…

? How could Grokking be made workable? Mechanistic Interpretability, like using SAEs (EleutherAI did), also methods Heretic employs to decensor, surely you can see some "modding" d…

It's way overblown... The "anti LLM" or "no AI" clause is really more a "covering all ass" clause. If you want to improve Forgejo and have that AI PR; you better clean it up of all…

No. 1. "Grokking" was shown on 4-digit modular arithmetic with a 1-layer transformer; this article extrapolates it to AGI and a $10B training run with exactly zero intermediate evi…

The screenshot of ol' claude closed code with that ascii table tells it all: Vapor AIware. As if it really would work like that. The noise added by the verbosity alone is not taken…

I agree. This isn't what people struggle with to be honest. With the limitations, right now it is sloppy unfortunately... Yet cool to release it nontheless.

This entire steganographically marking feature is prolly vibecoded like everything regarding claude code. They can try sure, but classic cat and mice game dictates: thankfully the …

Looks good. Won't ever buy a font though.

usage billing over the monthly plan when deepseek is over x25 cheaper?

Very, very early already with GPT-3. The simple fact that people will act on it and believe just because what they insinuated as a prompt and the answer being churned out on the sc…

DeepSeek rules. I'm using it to do stuff that's not too big in scope, because I still need to remain in charge. Even for this, western competitors have no chance, least Anthropic a…

Dev tools. The debugger is something for example that Microsoft ostensibly keeps to their own products, and how they totally slaughtered omnisharp. It killed my daily csharp vscode…

[flagged]

Thank you. Looking at the WebDAV in nginx, this is exactly what I searched for, wanted to read, and confirmed my suspicions ^^ But this one takes the cake truly... https://red.anth…

Where is the link to the advisories then? :/

Agreed, also amazing citations in the parent comment ^^

I don't buy it. A lot of stuff this finds is also just simply wrong, benignly reported as true, despite upper/lower layers in the code burying the possibility of a vulnerability ac…

You are absolutely right! Kidding, but the analogy sits comfortably with me. I wonder though if this kind of behavior is potentially harmful, most likely less than drugs but noneth…

The future of open washing

[dead]

I’d call it “open washing”, but it looks cool. Good luck with it

"The signal is not just stable. It also has high capacity." stopped reading right there also it's nothing that anybody using tails for example should have to worry about. Nothingbu…

depending how hard the "the brain is a muscle" saying applies, there is no way using LLMs/chatbot systems/AI is not going to deteriorate your brain immensely.

I worry about just becoming a product user. Some people push Claude and Claude Code at work, weights are closed, even Claude Code is completely closed and proprietary. If terms, wh…