Chatbots like ChatGPT can be jailbroken to output harmful text. But what about robots? Can AI-controlled robots be jailbroken to perform harmful actions in the real world? Our new paper finds that jailbreaking AI-controlled robots isn't just possible. It's alarmingly easy. 🧵

Collaborative AI runs on interactivity: machines and people, working in real time, across every modality. Solving it takes a community, join us.

We are offering grants of $100,000 Tinker credits to researchers advancing the field of human-AI interactivity. Submit your proposals by June 19th! thinkingmachines.ai/news/int…

Sharing our work on full-duplex multimodal models -- real-time interaction that's natural and intuitive without compromising on intelligence. We started Thinky in part to differentially advance capabilities for human-AI collaboration, which are underemphasized relative to intelligence/autonomy because they're harder to eval. In the future, we think every AI system will have something like an interaction model as the outer user-facing layer, continually keeping the user informed and learning what they actually want.

In the past few months, we had a lot of fun (and stress 😅) to produce 12 versions ( many subversions) and 137 pages in our training run log book. Turns out human-human collaboration is important to improving human-AI collaboration. 😊

Thinky's secret plan: 1: Increase Human<->AI bandwidth 2: Raise ceiling of human AI intelligence 3: Help humans continue as main-characters in the new world We are at Step 1. Interaction Models are great real-time collaborative tools for humans. Here's a preview:

Today we're sharing our work on interaction models. A new class of model trained from scratch to handle real-time interaction natively, instead of gluing it onto a turn-based one. youtu.be/A12AVongNN4

At @thinkymachines, we think models should talk with you, not at you. Check out our latest research preview: interaction models — built to actually hold a conversation.

Welcome @luke_drago_ and @LRudL_ to @thinkymachines. They started Workshop Labs determined to build AI that keeps the future human. They’ll continue that mission at Thinking Machines, where we create powerful AI systems that think alongside humans and extend our agency.  From Tinker to our research grants to the work we're doing to advance the frontier, everything we do is in service of the same mission -- AI that keeps our civilization empowered. Luke and Rudolf have been building toward the same thing. There's a path for AI to make humans matter more. Glad to have them working on it with us.

Workshop Labs is joining @thinkymachines. We believe there's a path for AI to make humans matter more. We couldn’t be prouder to join Thinking Machines to see this work through. workshoplabs.ai/blog/wsl-joi…

I found this to be a wonderful paper. I made a related observation last year: the hparam scalings for learning rate and batch size in our signSGD paper (from 2018!) have been found to be compute optimal in recent LLM scaling studies (1/4)

For the past two months, I've been working on Safety at @hark_labs. Today, we are coming out of stealth. We are building the most advanced personal intelligence in the world, and I want to say a little bit about how we plan to do this safely 🧵

If I had to compress my PhD into one idea, it is this "The data a model sees early in training leaves an imprint on its representations that is very hard to undo later" This thread runs through - Rephrasing the Web - Safety Pretraining - TOFU This is the Finetuner’s Fallacy🧵

Models are typically specialized to new domains by finetuning on small, high-quality datasets. We find that repeating the same dataset 10–50× starting from pretraining leads to substantially better downstream performance, in some cases outperforming larger models. 🧵

Building technologies for better human-AI collaboration on next gen hardware at scale. Exciting.

Grateful to Jensen and @nvidia team for their support. Together, we’re working to deploy at least 1GW of Vera Rubin systems, bringing adaptable collaborative AI to everyone. thinkingmachines.ai/nvidia-p…

We are partnering with @nvidia to power our frontier model training and platforms delivering customizable AI. thinkingmachines.ai/news/nvi…

I highly recommend this blog post from Nicholas Carlini on how to do great research:

AI is getting great at math, but how good is it at solving real research problems in areas outside of those covered by Erdős problems? Towards gauging this, I have started putting together a list of unsolved research problems in mathematical statistics and machine learning, sourced from recent papers in a leading statistics journal, the Annals of Statistics (with some bonus COLT open problems: solveall.org/. Currently >100 problems. In my view, much of the value of AI for researchers in the mathematical sciences stems from helping with their own research problems. These are problems without known solutions. There are many math benchmarks, but few with the following properties: (1) of a realistic research-level, so that solving them can potentially lead to a publication in a top journal (problems discussed in papers already, not contest math, not Millenium problems, not problems created for a benchmark, not problems that have a known solution); I'd say Erdős problems are the best example of this. (2) cover problems outside of the usual focus (combinatorics, number theory, ... ) of Erdős problems. Especially under-represented are domains of applied math, along with statistics, operations research, etc. I'm interested in statistics and ML, so that's where I started, but this could grow over time. Hope this can grow into something useful to the community! Happy to hear your thoughts...

LLM-enabled robots can cause physical harm in the real world. How do we safeguard them? Our new paper introduces RoboGuard, a safety guardrail for LLM-enabled robots — accepted to IEEE Robotics and Automation Letters (RA-L). 🧵