Boutique AI

Boutique AI! Customised AI, Recommender Systems, Bot tech!

[Andrej Karpathy @karpathy]

Dusting off this tweet from April because this gap in shared understanding of LLM capability is *widening*. It's now less "two groups of people speaking past each other" and more a sharp funnel.

- Napkin math somewhere around 6B people (~75% of the population) have barely come in contact with LLMs at all.
- Around 1-2B (~20%) are casual and infrequent users of free-tier ChatGPT-like products. This group sees derpy chatbots and treats them a bit like a better Google search, a writing aid, or etc. Maybe an agent tries to book you a flight. I have non-tech friends who (reasonably, imo) say they have not much use for it at all. Even many professionals outside of math&code are in this tier. For example, execs and many other functions spend a lot of their time talking to other people, so while they understand what is happening intellectually, it is still second-hand and a bit abstract.
- Now we get to professional use of frontier-grade LLMs in math&code. Somewhere around 20M people (0.2%) see first-hand that large, complex projects that used to take them weeks/months can now be completed by agents with a prompt. Building apps, copying apps, translating apps, decompiling apps from binaries... This has all happened very quickly and recently - less than 1 year ago, I was writing code manually by hand, typing memorized computer code commands into a code editor character by character, occasionally pressing Tab to autocomplete a little chunk of code.
- And finally we get to the ~5,000 people (~0.00006%) with access to frontier-grade systems internally. The external world has seen the preview. It looks like swarms of thousands of agents collaborating over weeks on software mega projects: minting zero days, running cyber attacks and defenses at machine speeds, discovering new science, advancing the frontier of mathematics. Things that would have taken top professionals in the industry years of work. Meanwhile, human review and comprehension are starting to fall behind. For example, people are still involved in the "archeology" of the OpenAI-HF incident from many months ago. Mathematicians may be poring over the 722 manuscripts on frontier mathematics for a while.

The funnel is driven by a combination of factors:
1. The impact scales with ambition, problem size, and horizon. A question with a paragraph answer barely stresses the system. You need a reservoir of big, difficult problems that you really care about. This aspect drives the consumer / professional dimension of the funnel.
2. The jaggedness of the system (which I have written about a lot separately). Capability peaks in domains that are digital, verifiable and economically valuable. This is because LLM capability emerges from reinforcement learning on verifiable rewards on a curated environment mixture driven by revenue potential. This aspect primarily drives the area (e.g. math&code) dimension of the funnel.
3. Access. Free-tier, paid-tier, internal.

So this is the weirdness of the moment. The general public has mostly not interacted with these systems. When they have, it looks like a derpy chatbot. The majority of professionals still see only a modest uplift. And a small sliver of professionals are experiencing the vertigo of the curve going vertical. And it is all happening at the same time.

[Diogo Almeida @CompleteSkeptic]

After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI?

I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev

• 20-200x faster
• 40-400x cheaper (w/ output tokens free)
• Frontier composable intelligence optimized for decisions

AFAICT the shortest path to AI-based economic revolution

[Mo Bavarian @mobav0]

This is a surreal moment. Few people could have predicted that the AI will advance to solving math problems at the highest level only a few years after GPT-2/3. The models then couldn't reliably solve grade school math problems. They barely were good enough to draft emails. They still very much felt like statistical parrots.

All things that looked like fundamental limitations slowly faded with some advances (e.g. high scale RL) in a span of a few years, which is really a short time. We should behold this moment both in awe and disbelief. What will a few more years of progress bring? How is it going to impact the world and society at large? Are we ready for the tsunami of intelligence at our fingertip?

More than any other moment, this feels to me like the eve of singularity. Glasswing & huggingface incident further increase the gravity.

A few years ago, deep down I felt working on alignment is premature. It's nice to do if it's your passion, but the shapes of things weren't clear enough for it to be critical in my opinion. The chances that you end up working on things that are useless for aligning the actual AGI was high. It's different now. Now, it feels like it's the most critical thing facing us.

https://t.co/SDdrvNRefr

[Kai Arulkumaran @kaixhin]

Brief review of recent deep learning work that has relations to concepts in symbolic AI - a broader context for ideas such as graph networks, disentangled representations, program induction with neural networks, etc.

[Stanford NLP Group @stanfordnlp]

Amid plenty of noise, if you read this whole thread 👇 there’s some worthwhile thoughts about what’s needed to incorporate compositional reasoning into modern (i.e., neural) artificial intelligence systems—including some explicit mention of #NLPproc.

[Andrej Karpathy @karpathy]

Hidden Technical Debt in Machine Learning Systems https://t.co/szYRTtSpDd on some of the new joys and struggles of deploying machine learning models in the wild. Still a long way to go to establish new language and design patterns for programming the 2.0 stack

[Christopher Manning @chrmanning]

Just a couple of years ago, I found it hard to believe that vision people were still all working with rectangular bounding boxes. I guess they’ve fixed that now. 🙂

[Founders Factory @foundersfactory]

Sunday long read: are we speeding towards #AI consciousness? Founders Factory Chief Scientist @jeffreynsk explains https://t.co/fc9Mc2E6z3

[Jeffrey Ng @jeffreynsk]

Social and Emotional Intelligence will become more important as robots become more technically intelligent and AI Innovation happens. #UIPathForward #FutureOfWork #McKinsey #JacquesBughin

[Jeffrey Ng @jeffreynsk]

Future of work: Businesses are looking to adopt #AI to retain a competitive advantage rather than displacing their workforce. #AI_Innovation

[Farza 🇵🇰🇺🇸 @FarzaTV]

I'm happy to finally release DeepLeague. The first combination of computer vision, deep learning, and esports analytics + a dataset of over 100k labeled images to drive AI / esports research forward. Completely open-source. Enjoy. https://t.co/OW691Se8VV

[hardmaru @hardmaru]

Reinforcement Learning for Improving Agent Design: What happens when we let an agent learn a better body design together with learning its task?
article: https://t.co/QsGesXVuOD
pdf: https://t.co/NZPYPIBInA

[Boutique AI @AI_Reki]

@KaiLashArul Or a Deep Net that decides what to do based on the website that you point it at, e.g. order me some pizzas!

[Artificial Intelligence @AITechmeme]

Making Sense Of, And Making Progress With, Industry 4.0 via @forbes https://t.co/DaoBC0MiTh

[Nando de Freitas @NandoDF]

Machine Learning Confronts the Elephant in the Room. John Tsotsos on human vision. https://t.co/wU8QpOYuHB

[Nando de Freitas @NandoDF]

Deep reinforcement learning to make basketball video games look more realistic https://t.co/F2AbkqOJus

[Nando de Freitas @NandoDF]

This is such a cool visualisation, hammering a fact that is often overlooked: that brains spread through the body, and are not caged in the head. https://t.co/DBr9P9fwxw

[Neal Khosla @nealkhosla]

This is really cool. Interpretable ML models that self-learn representations (don’t need feature engineering) and have good model capacity are going to be so important going forward.

Note their finding that "optimal interpretable features don't equate with human intuitions."

[Yann LeCun @ylecun]

Unsupervised pre-training is starting to work quite well for learning visual representations.

Here is a new paper from FAIR-Paris that demonstrates how this works. https://t.co/MKPqdPRs05

[Boutique AI @AI_Reki]

Congratulations to @riedelcastro and the @BloomsburyAI team for the acquisition! https://t.co/EGuhvDEmLV

[Yann LeCun @ylecun]

It's official!
Bloomsbury AI joins Facebook.
Welcome Guillaume Bouchard, Sebastian Riedel, and the team! https://t.co/ajTH0CjhpV