“@OpenAI Groq’s recent hardware breakthroughs have been going viral on X. Groq (not Grok) uses LPUs instead of GPUs, allowing the chatbot to run LLMs at nearly instantaneous response times. This unlocks a whole new world of potential AI and user experiences. https://t.co/EnJnd3jEQm” / X – https://twitter.com/rowancheung/status/1760170171002438096

“Seeing as I published my Tokenizer video yesterday, I thought it could be fun to take a deepdive into the Gemma tokenizer. First, the Gemma technical report [pdf]: https://t.co/AgxjBJfh0T says: “We use a subset of the SentencePiece tokenizer (Kudo and Richardson, 2018) of…” / X – https://twitter.com/karpathy/status/1760350892317098371

“Fun LLM challenge that I’m thinking about: take my 2h13m tokenizer video and translate the video into the format of a book chapter (or a blog post) on tokenization. Something like: 1. Whisper the video 2. Chop up into segments of aligned images and text 3. Prompt engineer an LLM…” / X – https://twitter.com/karpathy/status/1760740503614836917

“Sora from OpenAI and Gemini 1.5 from Google are both super impressive. Both rely on long context transformers. I think the lesson is that semi parametric models are great – condition on all past data, whether real or self generated, just like a good Bayesian :)” / X – https://twitter.com/sirbayes/status/1758377392928899410

LLM evaluation at scale with the NeurIPS Efficiency Challenge – https://blog.mozilla.ai/exploring-llm-evaluation-at-scale-with-the-neurips-large-language-model-efficiency-challenge/

I worry our Copilot is leaving some passengers behind – Josh Collinsworth blog – https://joshcollinsworth.com/blog/copilot

“The mobile S-curve ends, and the AI S-curve begins There’s never been a bigger contrast between mobile and AI — it’s the end of one technology curve, and the start of the other. It’s been 15 years since the App Store was launched; while the generative AI revolution started… https://t.co/5UlHYcmhJc” / X – https://twitter.com/andrewchen/status/1760698184966504475

Introducing DatologyAI — Making models better through better data, automatically – https://www.datologyai.com/post/introducing-datologyai-making-models-better-through-better-data-automatically

How to jointly tune learning rate and weight decay for AdamW – Fabian Schaipp – https://fabian-sp.github.io/posts/2024/02/decoupling/

[2402.10422v1] Pushing the Limits of Zero-shot End-to-End Speech Translation – https://arxiv.org/abs/2402.10422v1

[2401.18079] KVQuant: Towards 10 Million Context Length LLM Inference with KV Cache Quantization – https://arxiv.org/abs/2401.18079
My benchmark for large language models – https://nicholas.carlini.com/writing/2024/my-benchmark-for-large-language-models.html

Be Sure To Read “This Week In AI”

This week’s executive overview and top links are here:

AI News #21: Week Ending 02/23/2024 with Executive Summary and Top 36 Stories

The post you just read is an extension of my weekly newsletter, This Week In AI, an executive summary of the top things to know in AI. Each week, I create an accessible overview for laypeople to feel confident they are conversant with the week’s AI developments. I include a curated list of must-click links of the week, to offer everyone a hands-on opportunity to explore the most intriguing updates in artificial intelligence across various categories, including robotics, imagery, video, AR/VR, science, ethics, and more. Beyond the overview, I post these topic-based deeper dives (below). If you haven’t read this week’s overview, I recommend starting there.

Credits/Sources

Most of these weekly links come from just a few prolific oversharing sources. Please follow them, as they work hard to find the news each week and they make it a lot easier for me to compile.