hislaziness

This is amazing, now waiting for someone to run an llm on the cpu. Eventually it will be LLMs all the way down...

Am I missing something here. Not finding a major bug/vulnerability just means that maybe the code is really good, not that the model is not what is claimed?

Would it be more appropriate to compare LLMs to Autotunes rather than pianos?

Akamai CTO Robert Blumofe offers four tips for business leaders striving to foster AI fluency by empowering employees with the right tools and best use cases.

useful insights but definitely a PR article.

Great. I enjoy these kind of articles. My all time favorite book for 'C' is Expert C Programming: Deep C Secrets.

Terence Tao on the matter - https://imgur.com/a/terence-tao-on-supposed-gold-imo-sMKP0bm

As I understand, the LLM uses the techniques of searchformer - https://arxiv.org/abs/2402.14083. To do "slow thinking" doing a A* search using a transofrmer.

It is not just MSRP, management and operations cost too. The article goes into the details of this.

The details are in the article. They have done the math.

TLDR: Don’t buy H100s. The market has flipped from shortage ($8/hr) to oversupplied ($2/hr), because of reserved compute resales, open model finetuning, and decline in new foundation model co’s. Rent instead.

Is the AI infra bubble already bursting?

The screen seems to be stuck at Wait a second... for me.

I tried a few local LLMs. None of them could give me the right answer for "How many 'r's in straberry. All LLMs were 8-27B.

I know you mean this in jest, but we are much closer to this than we would imagine, the use of LLMs to process communication / translation is becoming ubiquitous. We are 1 bad translation away from a disaster.

I also use some email providers ability to have +xyz at the end of the username. So for a registration I would for [email protected]. has helped me track spam and leaks in the past.

Cool. I will try it out. I tried the same with ollama, the non english part needs a lot more polish. Do you see the outcome being any different?

This is pretty awesome. How did you do it? Any blog on the detials?

The model description on huggingface says - Model size - 12.2B params, Tensor type - BF16. Is the Tensor type different from the training param size?

I just checked huggingface and the model files download is about 25GB but in a comment below someone mentioned it is 8fp quantized model. Trying to understand how the quantization affects the model (and RAM) size. Can someone please enlighten.

isn't it 2 bytes (fp16) per param. so 7b = 14 GB+some for inference?

The title does not do justice to the article. It talks about a new classification system that OpenAI has introduced for LLMs

Same here. Coincidentally I joined in June 2009 and this is the only social media I indulge in.

The articles states the case is for apple's monopoly in the premium phone market. Yet it compares the market share with low cost android phones.

I have experience using the Motofone - https://en.wikipedia.org/wiki/Motorola_Fone . It was not the best of experience especially due to the lag which many have mentioned here. E-Ink displays have come a long way, hoping this time it is a better experience.

Inspired by MalbolgeLISP which was a lisp written in Malbolge.

Came across this book today. Good introduction and deep dive into ML.

As a customer you could propose additional packages, send the request via the support channel...