A picture is worth … less than a word?

Call for Papers 📣

Working on AI and sustainability?
Join our IJCAI 2024 workshop on Sustainable Transition with AI❗️

We invite research not only on climate change but any sustainability topics!

⏰Submission: 7 May
🌏stai.jeju.ai

#IJCAI2024 Christopher Manning Y.J. Lee

I'm ecstatic to share that preorders are now open for the AI Snake Oil book! The book will be released on September 24, 2024.

Arvind Narayanan and I have been working on this for the past two years, and we can't wait to share it with the world.

Preorder: princeton.press/awecv19z

Why would any company release a strong LLM for free?

This is ServiceNow (mkt cap ~$160Bn) stock price since the release of StarCoder. Only required a small amount of compute and a handful of people while building up a lot of valuable know-how fast.

It's not a zero-sum game!

If this were a science paper, you would expect a country that picks its science workforce at random as a “weak baseline” and a leading nation like the US to actively experiment towards state-of-the-art, or at least beat the baseline.

Not providing a guaranteed path for…

Who knew a class could turn into a family reunion? 😮 Professor Chris Manning Christopher Manning & Dan Jurafsky Dan Jurafsky just found out they're 8th academic cousins in CS324H! Talk about #ScholarlyRoots ! #StanfordLife
#NLP #StanfordAI #StanfordNLP #NaturalLanguageProcessing

Technologists often ask, “How can we replace humans doing existing tasks?”
But they should ask, “What entirely new thing can we now do that we’ve never done before?”

From my piece for the American Enterprise Institute 'Beyond the Turing Test: Harnessing AI to Create Widely Shared Prosperity.'…

rime.ai/blog/filler-wo…

ummmmmmm

It looks like the behavioral psychology folk still have some work to do to get beyond the replication crisis….

“HAI was the first institution in the public sector that was set up devoted not only to AI innovation, but also to engaging policy, industry, and civil society to ensure this technology is developed with humans at the center,” says Fei-Fei Li. stanford.io/3TT5SNS

Big investments happen due to: ① clear breakthrough results (AlexNet), ② crisis (atom bomb), ③ politics (Cycorp).

Otherwise small studies bubble along; that was all of probabilistic AI, neural nets, diffusion models & mRNA vaccines. Neurosymbolic AI so far fails all three.

I am pretty surprised by the lack of urgency among many (but not all) academics in addressing what seems to be one of the biggest developments in modern times - we accidentally built a machine that produces something that looks like language & thought

Why? What does it teach us?

A big problem with AI is if you don't want it to work for something it won't work. If you want it to work for something, you often (but not always) can make it work with some prompting and experience.

It leads to stark divisions in how people see AI.

Crypto money secretly funding AI doomerism, lobbying for AI regulations, and effectively working against open source AI platforms.

The most common question I've gotten since releasing RewardBench is: how do you use a DPO reward model as a ranking model?

It's quite simple, the numbers are just less intuitive than with classifiers.
1. Take the reward directly from the DPO paper.
2. Take existing DPO code,…