gardnr

12 ms·

1. 211.▲
   by gardnr 2y ago

   For those readers wondering if “ C Structures for the Down and Out” is a real book: please be patient. Claude is writing it as we speak.
2. 212.▲
   by gardnr 2y ago

   Have you been to your local library? Hah! It’s a bit different but kind of the same. LibGen manifests the idea that humanity and its progress are more important than copyright. I wonder if Anthropic, OpenAI, or Meta.ai have spent much time
3. 213.▲
   by gardnr 2y ago

   I'll check it out. There is also LlocalSearch: https://news.ycombinator.com/item?id=39923404
4. 214.▲
   by gardnr 2y ago

   Edit: I just found this. I'll give it a try today: https://github.com/0ssamaak0/SiriLLama --- Open WebUI has a voice chat but the voices are not great. I'm sure they'd love a PR that integrates StyleTTS2
5. 215.▲
   by gardnr 2y ago

   I was surprised: https://arxiv.org/abs/1912.09363 It's not exactly an LLM but it works in a similar fashion.
6. 216.▲
   by gardnr 2y ago

   Kia kaha!
7. 217.▲
   by gardnr 2y ago

   Parent were referring to https://en.wikipedia.org/wiki/Natural_person It's legalese for a "human being".
8. 218.▲
   by gardnr 2y ago

   In the late 90s, Borland's installation support phone line would tell people to turn their computer on its side 90 degrees so that the CD-ROM would read the installation media. It usually worked.
9. 219.▲
   by gardnr 2y ago

   If you don't `npm config set ignore-scripts true` then you are basically executing random unknown code every time you do an `npm install`. Modules can change hands and become malicious, which is much more plausible when there are thous
10. 220.▲
    by gardnr 2y ago

    The linked GitHub issue states: > I created a new project using crewai==0.35.8 > At this site: https://docs.crewai.com/telemetry/Telemetry/ > > It's pivotal to understand that NO data is collected
11. 221.▲
    CrewAI collects prompts via telemetry, can't be disabled(github.com)

    2 points by gardnr 2y ago|1 comments
12. 222.▲
    by gardnr 2y ago

    Not enough info. Do they want near-realtime responses? Will they all hit it at the same time? Can you put some workloads in an overnight batch queue?
13. 223.▲
    by gardnr 2y ago

    It looks great! Have you considered running it through https://www.quilter.ai/ before sending it out? It's free. You can hear more about it in this podcast: https://wandb.ai/site/resources/pod
14. 224.▲
    by gardnr 2y ago

    SST is working to relase ion which uses pulumi instead of CDK: https://sst.dev/blog/moving-away-from-cdk.html
15. 225.▲
    by gardnr 2y ago

    Thanks for that. It's late here. I'll go to sleep now.
16. 226.▲
    by gardnr 2y ago

    https://archive.is/R5zkq It is relevant to point out that NYT is engaged in a legal battle with OpenAI. This article refers to "AIs like ChatGPT" which clearly shows who it is targeted at. For people thay don&#x27
17. 227.▲
    by gardnr 2y ago

    Yi is led by Dr. Kai-Fu Lee. They have been releasing a lot of really good models over the last ~6 months. Their previous (1.0?) Yi-34B-Chat model ranks similar to GPT-3.5 on Chatbot Arena. [1] A quantized version of that model can be run o
18. 228.▲
    by gardnr 2y ago

    Got that one backwards: > Hyperventilation is irregular breathing that occurs when the rate or tidal volume of breathing eliminates more carbon dioxide than the body can produce. https://en.wikipedia.org/wiki/Hyperve
19. 229.▲
    by gardnr 3y ago

    Hah! You are correct. Sorry for the confusion.
20. 230.▲
    by gardnr 3y ago

    The MI300X has more that twice the memory and is 25-50 the cost of an H100. It's a great deal if they can make it work. It could be great for everyone else that TensorWave is willing to invest in it and hopefully help drive improvement
21. 231.▲
    by gardnr 3y ago

    I prefer adb hell. https://stackoverflow.com/questions/18097790/adb-hell-comman...
22. 232.▲
    by gardnr 3y ago

    The first time I went to set up kubernetes the documentation explained what each of my three machines would do. It's different now but I still find the syntax needlessly verbose. So now instead, I have a rediculously long docker compos
23. 233.▲
    by gardnr 3y ago

    Here is a description: https://js.langchain.com/docs/use_cases/query_analysis/techn... > When a user asks a question there is no guarantee that the relevant results can be returned with a single query. Som
24. 234.▲
    by gardnr 3y ago

    > By using lithium-iron-phosphate as the cathode material, BYD can make the batteries much cheaper. How is LiFePO₄ "next-gen"?
25. 235.▲
    by gardnr 3y ago

    I like the way you think but 2b might prevent that.
26. 236.▲
    by gardnr 3y ago

    > We have found infringing material in your website which indeed is OTHER 'Plex Server'. It would appear Plex are claiming there was 'Plex Server' in the repo.
27. 237.▲
    by gardnr 3y ago

    You probably want to build a retrial augmented generation pipeline. If you do end up wanting to fine tune then use qlora with axolotl or unsloth to prove your hypothesis on a smaller model and then evaluate if you want the marginal gains yo
28. 238.▲
    by gardnr 3y ago

    > To mitigate such errors, we propose designating functions as unique functional tokens. I just skimmed the paper but this seems to be the crux of it. They map functions to a single token and can then fine-tune models to use the token in
29. 239.▲
    by gardnr 3y ago

    Hotdog or not?
30. 240.▲
    by gardnr 3y ago

    $2B is a low valuation for a company that ostensibly can replace most software developers.