This chart shows the three big questions of AI right now (even given the limitations of MMLU as a test):
1) OpenAI was very far ahead, now it isn’t. Will GPT-5 be another leap?
2) Will open source models converge with closed source?
3) Will rapid progress continue or hit a wall?

It's been a wild ride. Just 20 of us, burning through thousands of H100s over the past months, we're glad to finally share this with the world! 💪

One of the goals we’ve had when starting Reka was to build cool innovative models at the frontier. Reaching GPT-4/Opus level was a…

The way a lot of ppl seem to see the world is:

- if things are getting bad, that's news
- if things have been bad but are getting better, that's also bad news
- if things are good, that's boring so it's not news

The problem with this approach isn't just that you persuade…

“his research analysts could now wake up two hours later than usual, because A.I. can read overnight economic data and create a written draft of analysis to work from.” nytimes.com/2024/04/10/bus…

“But the consulting giant Accenture estimated that A.I. could replace or supplement nearly three-quarters of bank employees’ working hours across the industry.” nytimes.com/2024/04/10/bus…

Amazing work. People have been working on this pre-GPT but the simplicity will be a big deal for on-ramping less technical folks. Will take it for a spin with my 'humanities' glasses on.

Quick writing tip: if you're like me and have a complicated brain, then just cut your first paragraph and start there. I always back into arguments. Somewhere around para. 2/3 the insight emerges!

I found a guy who goes around the US asking 70 to 100-year-olds their:

• Biggest Regrets
• Biggest Lessons
• Advice to their younger self

Oddly enough, most of their answers are the same…

Here are my top 8:

I like this. I've always said we are the squirrels of evolution. Playful, but convinced everyone is out to get us.

This is by far the best non-technical Natural and Artificial Intelligence book anyone could read. This comprehensive, well-researched, crisply clear, sharply focused and illuminating book is a thing of beauty. It is the book I wish I had had when I started my AI career 30 years…

'There is no single best way to prompt AIs (researchers keep discovering new techniques and approaches), making trying to show that AI can do something dependent on a mix of art, skill, and motivation.' Motivation is the number one ingredient. open.substack.com/pub/oneusefult…

We have come to learn a lot about how we learn, but students often rely on outdated study techniques.

Here's the results of a large meta-analysis on learning. Skip highlighting, summarizing & rereading. Instead, practice, quiz & explain (AI can quiz you). journals.sagepub.com/stoken/rbtfl/Z…

“I worry that the current obsession with mental health awareness is disempowering parents from helping their adult children handle ordinary things.” nytimes.com/2024/04/08/opi…

'Literally I got 70 out of 70 in only 6 minutes.' We're going to need to up our assessment game. Again. tiktok.com/@answersai.com…

Everyone is talking about addressing GPTs effect on classroom writing. But what about reading? You really don’t need to read anything anymore if you don’t want to.

These takes are a pass. The fast forward to AI doing all our work overlooks all the work and pollution it would take to get there. Like most AI reporting waste of ink. nytimes.com/2024/04/05/opi…

“How did this happen, and can it possibly be worth it?” As someone inside higher Ed: it’s not. nytimes.com/2024/04/05/you…

While I'm loving the 3 Body Problem, warts and all (Goonies!), the one thing they really get wrong is that no one does any work! If you plop an academic down with time to kill they will absolutely work! Read, math, outline a paper, whatever. We don't sit still.