Viewing profile — jpdus

germany as well. Claude down too

This is also confirmed by internal cline statistics where Opus and Gemini 2.5 pro both perform worse than Sonnet 4 in real-world scenarios https://x.com/pashmerepat/status/19463924…

My comment from the original submission [1]: --- As someone who is in general skeptical of programs like this (and an European) there are 2 remarkable / timely things about this: -…

I think no one believes that R1 costs $5.5m from scratch. People in this project (most, not all) are very aware of the realities in training and are very well connected in the US a…

I agree that the announcement should´ve talked more about goals and performance than regulatory stuff ;-). But I think there is a new understanding among the bureaucracy that regul…

There definitely is - but that we, as a startup that is barely a year old and not widely known outside our niche in AI dev circles and on Huggingface, are part of this is already a…

might be debatable - but I tend to agree with Dario Amodei on this; my guess is that R1 is 7-10 months behind the internal frontier at the big labs, while having a few small novel …

As someone who is in general skeptical of programs like this (and an European) there are 2 remarkable / timely things about this: - This project doesn't just allocate money to univ…

ellamind | Software Engineers (Full stack/AI/SRE) / Chief of Staff| Full-time | On-Site / Hybrid /Remote Bremen, GERMANY| https://ellamind.com I'm Jan, the Co-Founder of ellamind a…

Wow, actually this cookbook is really bad? I expected something like the OpenAI or Anthropic cookbooks, but this seems to be some AI generated low-quality content without any code …

This is such a good comment and should be auto-posted to every pro-nuclear thread. I get why people want to believe in nuclear and i'd wish that we invested a lot more in developme…

I have the same question. Noticed that Ollama got a lot of publicity and seems to be well received, but what exactly is the advantage over using llama.cpp (which also has a built-i…

Hey, imho best overall technical intro to LLMs (I guess that´s your main interest as you mentioned qlora + llama) is by Simon Willis [1]. Additionally or if you prefer videos, the …

Does nobody question how they get from a non-preregistered, 9-mouse study (where the treatment group gets only 6%-10% acoholic drinks and nothing non-alcoholic for 10 weeks) to thi…

We now have a (experimental) working HF version here: https://huggingface.co/DiscoResearch/mixtral-7b-8expert

Don't switch to FF if you're a heavy user. I am on Firefox as my main browser for web and mobile since 4 years. I am just in the process of switching back to Chrome, as Firefox got…

It isn't. Compared to the original Orca model and method which spawned many of the current SotA OSS models, Orca 2 models seem to perform underwhelming, below outdated 13b models a…

For other (non-code) benchmarks, people are having the opposite experience: "I benchmarked on SAT reading, which is a nice human reference for reasoning ability. Took 3 sections (6…

For P1 (which is only to prove safety) this may be the case, but for pivotal studies in P2 and later, participants are almost always randomized so cherry picking shouldn't be possi…

EY had an interesting take on this: "Here we are, exploiting the shit out of the equivalent of naïve six year olds working online, forcing kindness and sympathy to be removed from …

Care to share your detailed stack and command to reach 50t/s? I also have a 7950 with DDR 5 and I don't even get 50 t/s on my two RTX 4090s....

Either you only use the GPU sporadically or your math is very different from Tim Dettmers` [1]: "The break-even point for a desktop vs a cloud instance at 15% utilization (you use …