Palmik

18 ms·

301.
Anthropic Debuts Claude 3(cnbc.com)
5 points by Palmik 3y ago|0 comments
302.
by Palmik 3y ago
There's no magnet link for this new (Mistral Next) model as of right now.
303.
by Palmik 3y ago
It's just as likely that Google knew OpenAI has their announcement planned for today and wanted to preempt it. Happens all the time.
304.
by Palmik 3y ago
I'm very uncomfortable with this technology being accessible only to a small and arbitrary subset of the population.
305.
by Palmik 3y ago
Do you have a source on Mistral API, etc. being based on TensoRT-LLM? And what are the main distinguishing features? What I like about vLLM is the following: - It exposes AsyncLLMEngine, which can be easily wrapped in any API you'd lik
306.
by Palmik 3y ago
This is a dupe, see here for more discussion https://news.ycombinator.com/item?id=39300679
307.
by Palmik 3y ago
Apparently it's just GPT 3.5 which is freely accessible anyway. If it was GPT 4 / ChatGPT Plus, I would understand.
308.
by Palmik 3y ago
Except people modify "non open source" but "weights available" models all the time. In fact, this very model is such modification (fine tune) of the original base model.
309.
by Palmik 3y ago
> Yes, you could try convincing a very good friend or family member to lend you their street address if you really can't or won't pay for a place of residence yourself. In what way is that leeching? I would have no problem lett
310.
by Palmik 3y ago
The effort started at the ~beginning of February and ended at the ~end of October [1]. The dataset you link is from April and had "unsafe" content filtered. [1] https://m.youtube.com/watch?v=gqtmUHhaplo&feature
311.
by Palmik 3y ago
Tangential, but why didn't the OpenAssistant team (lead by the author of the video) release the OpenAssistant dataset? As far as I know, the project was shut down, and only some initial highly filtered version of the data got released.
312.
The Pixel 9 might come with exclusive "Pixie" AI assistant(arstechnica.com)
2 points by Palmik 3y ago|0 comments
313.
Audiobox: Meta's new foundation research model for audio generation(audiobox.metademolab.com)
2 points by Palmik 3y ago|0 comments
314.
by Palmik 3y ago
> we have NOT evolved where we are today by generating inputs to ourselves Sure, we have not evolved this way in the strictly biological sense, but we did greatly extend our capabilities this way. The jump in capabilities in the past ten
315.
LLM Arena. Mistral-small best open model. Gemini Pro beaten by 2 open models(huggingface.co)
9 points by Palmik 3y ago|1 comments
316.
by Palmik 3y ago
This is a naive take. How do you think Google collects or collected data for their safe-search classifiers? Now that's a sludge. Or how do you think Google evaluates search-ranking changes (or gather data for training various ad-rankin
317.
by Palmik 3y ago
The cost of hiring humans to write the trillions of tokens they trained from scratch would surely be much larger than the training cost. Except they avoided that cost by using what's available on the Internet. [1] Similarly, people are
318.
Gemini Pro, Mixtral (Mistral-Small) vs. GPT3.5 in LLM Arena(twitter.com)
2 points by Palmik 3y ago|0 comments
319.
by Palmik 3y ago
This looks amazing, well done. The pricing is very reasonable. (1) Seriously considering switching to this from my homebrew setup. Extension support (pg vector) would be great. (1) Looking at https://www.ubicloud.com/docs&#x
320.
Diffusion Models for Robotics Control(diffusion-policy.cs.columbia.edu)
75 points by Palmik 3y ago|13 comments
321.
Steering at the Frontier: Extending the Power of Prompting(microsoft.com)
2 points by Palmik 3y ago|0 comments
322.
AI image platform CivitAI dropped by cloud provider after reports of CSAM(engadget.com)
2 points by Palmik 3y ago|0 comments
323.
Anthropic Claude for Sheets(workspace.google.com)
1 points by Palmik 3y ago|0 comments
324.
by Palmik 3y ago
This is extremely impressive if benchmarks translate to real-world performance [1]. The mistral-medium beats GPT3.5 and also Gemini Pro (Google's best available model) with a huge margin on all available comparable benchmarks: https:&
325.
by Palmik 3y ago
Interesting interview. Cover Mistral's stance on safety and their business model plans. For safety: Keep releasing unencumbered pre-trained models, but offer tools to gain control over input / output safety. Business model: Cost e
326.
Interview with Arthur Mensch, CEO Mistral AI [video](youtube.com)
2 points by Palmik 3y ago|1 comments
327.
by Palmik 3y ago
Beyond what others said, I think this is an extremely impressive showing. Consider that their efforts started years behind Google's, and yet their relatively small model (they call is mistral-small, and also offer mistral-medium) is be
328.
by Palmik 3y ago
Inference should be closer to llama 13b, since it runs 2/8 experts for each token.
329.
by Palmik 3y ago
There are many free GPT frontends, especially web frontends, e.g. https://github.com/yakGPT/yakGPT
330.
Selecting for Intelligence in Dogs to Produce Human-Level IQ in 100 Generations(figshare.com)
3 points by Palmik 3y ago|1 comments

More ›