Wei Yang

[Wei Yang @weiyangedward]

@VIVOBAREFOOT I am still waiting for my refund. This company denies receipt of the package I returned more than a week ago. Customer service is useless.

[Wei Yang @weiyangedward]

@VIVOBAREFOOT I am still waiting for my refund. This company denies receipt of the package I returned more than a week ago. Customer service is useless.

[Wei Yang @weiyangedward]

@VIVOBAREFOOT I am still waiting for my refund. This company denies receipt of the package I returned more than a week ago. Customer service is useless.

[Wei Yang @weiyangedward]

@VIVOBAREFOOT I am still waiting for my refund. This company denies receipt of the package I returned more than a week ago. Customer service is useless.

[Tanishq Mathew Abraham, Ph.D. @iScienceLuvr]

NEW BLOG POST:

Debunking DeepSeek Delusions

(link in reply)

I discuss 5 main myths I have seen spreading online:

1. DeepSeek came out of nowhere
2. They're lying about the training cost
3. DeepSeek is extremely bearish for NVIDIA
4. DeepSeek are just copying American companies
5. DeepSeek is “sucking knowledge” from ChatGPT

[Tanishq Mathew Abraham, Ph.D. @iScienceLuvr]

Okay so this is so far the most important paper in AI of the year

[Aran Komatsuzaki @arankomatsuzaki]

Large Language Models are Effective Text Rankers with Pairwise Ranking Prompting

Proposes PRP, which uses the query and a pair of documents as the prompt for LLMs to perform ranking tasks.

https://t.co/sc6f0rMOB5

[Nature Portfolio @NaturePortfolio]

A study in @Nature shows that AlphaDev, an artificial intelligence agent from @DeepMind, discovered new and improved computer sorting algorithms widely used in C++ (a common computer programming language) libraries. https://t.co/enjbpTWy3m

[Tengyu Ma @tengyuma]

Adam, a 9-yr old optimizer, is the go-to for training LLMs (eg, GPT-3, OPT, LLAMA).

Introducing Sophia, a new optimizer that is 2x faster than Adam on LLMs. Just a few more lines of code could cut your costs from $2M to $1M (if scaling laws hold).

https://t.co/GrMY600lLO 🧵⬇️

[João Gante @joao_gante]

Text generation is the cornerstone of many NLP tasks, and 🤗 Hugging Face has cool news for you!

Instead of announcing a new LLM, today we're announcing a new text generation method in 🤗 transformers that is compatible with nearly ALL existing models: Contrastive Search!

1/

[OpenAI @OpenAI]

We’re rolling out web browsing and Plugins to all ChatGPT Plus users over the next week! Moving from alpha to beta, they allow ChatGPT to access the internet and to use 70+ third-party plugins. https://t.co/t4syFUj0fL

[Anthropic @AnthropicAI]

Introducing 100K Context Windows! We’ve expanded Claude’s context window to 100,000 tokens of text, corresponding to around 75K words. Submit hundreds of pages of materials for Claude to digest and analyze. Conversations with Claude can go on for hours or days.

[Paul Couvert @itsPaulAi]

ChatGPT has now a big problem.

Google just updated its free competitor, Bard.

Here are 8 things impossible on ChatGPT but that Bard can do (for free):