Daswin de Silva

The industry is rife with AI non-events that were wildly promoted and highly anticipated, but failed to deliver. @dswnds @latrobe theconversation.com/in-a-sea-of-hy…

Congrats to @ThankYourNiceAI, @sj_manning, @manlikemishap and @danielrock for their careful work on AI's labor market effects!

More autonomous AI systems that can use tools and work in teams are becoming increasingly common. @dswnds @latrobe theconversation.com/ai-agents-are-…

Nice - my AI startup school talk is now up! Chapters: 0:00 Imo fair to say that software is changing quite fundamentally again. LLMs are a new kind of computer, and you program them *in English*. Hence I think they are well deserving of a major version upgrade in terms of software. 6:06 LLMs have properties of utilities, of fabs, and of operating systems => New LLM OS, fabbed by labs, and distributed like utilities (for now). Many historical analogies apply - imo we are computing circa ~1960s. 14:39 LLM psychology: LLMs = "people spirits", stochastic simulations of people, where the simulator is an autoregressive Transformer. Since they are trained on human data, they have a kind of emergent psychology, and are simultaneously superhuman in some ways, but also fallible in many others. Given this, how do we productively work with them hand in hand? Switching gears to opportunities... 18:16 LLMs are "people spirits" => can build partially autonomous products. 29:05 LLMs are programmed in English => make software highly accessible! (yes, vibe coding) 33:36 LLMs are new primary consumer/manipulator of digital information (adding to GUIs/humans and APIs/programs) => Build for agents! Thank you again for the invite @ycombinator and congrats again on an awesome events! I'll post some links/references in the reply.

We're missing (at least one) major paradigm for LLM learning. Not sure what to call it, possibly it has a name - system prompt learning? Pretraining is for knowledge. Finetuning (SL/RL) is for habitual behavior. Both of these involve a change in parameters but a lot of human learning feels more like a change in system prompt. You encounter a problem, figure something out, then "remember" something in fairly explicit terms for the next time. E.g. "It seems when I encounter this and that kind of a problem, I should try this and that kind of an approach/solution". It feels more like taking notes for yourself, i.e. something like the "Memory" feature but not to store per-user random facts, but general/global problem solving knowledge and strategies. LLMs are quite literally like the guy in Memento, except we haven't given them their scratchpad yet. Note that this paradigm is also significantly more powerful and data efficient because a knowledge-guided "review" stage is a significantly higher dimensional feedback channel than a reward scaler. I was prompted to jot down this shower of thoughts after reading through Claude's system prompt, which currently seems to be around 17,000 words, specifying not just basic behavior style/preferences (e.g. refuse various requests related to song lyrics) but also a large amount of general problem solving strategies, e.g.: "If Claude is asked to count words, letters, and characters, it thinks step by step before answering the person. It explicitly counts the words, letters, or characters by assigning a number to each. It only answers the person once it has performed this explicit counting step." This is to help Claude solve 'r' in strawberry etc. Imo this is not the kind of problem solving knowledge that should be baked into weights via Reinforcement Learning, or least not immediately/exclusively. And it certainly shouldn't come from human engineers writing system prompts by hand. It should come from System Prompt learning, which resembles RL in the setup, with the exception of the learning algorithm (edits vs gradient descent). A large section of the LLM system prompt could be written via system prompt learning, it would look a bit like the LLM writing a book for itself on how to solve problems. If this works it would be a new/powerful learning paradigm. With a lot of details left to figure out (how do the edits work? can/should you learn the edit system? how do you gradually move knowledge from the explicit system text to habitual weights, as humans seem to do? etc.).

The encoder of a transformer, as explained by @rfeers.

I am pretty happy with this 30-minute summary of my views on the current state of AI and alignment. youtube.com/watch?v=w177Ov…

awards.acm.org/about/2024-tur… Machines that learn from experience were explored by Alan Turing almost eighty years ago, which makes it particularly gratifying and humbling to receive an award in his name for reviving this essential but still nascent idea.

New 3h31m video on YouTube: "Deep Dive into LLMs like ChatGPT" This is a general audience deep dive into the Large Language Model (LLM) AI technology that powers ChatGPT and related products. It is covers the full training stack of how the models are developed, along with mental models of how to think about their "psychology", and how to get the best use them in practical applications. We cover all the major stages: 1. pretraining: data, tokenization, Transformer neural network I/O and internals, inference, GPT-2 training example, Llama 3.1 base inference examples 2. supervised finetuning: conversations data, "LLM Psychology": hallucinations, tool use, knowledge/working memory, knowledge of self, models need tokens to think, spelling, jagged intelligence 3. reinforcement learning: practice makes perfect, DeepSeek-R1, AlphaGo, RLHF. I designed this video for the "general audience" track of my videos, which I believe are accessible to most people, even without technical background. It should give you an intuitive understanding of the full training pipeline of LLMs like ChatGPT, with many examples along the way, and maybe some ways of thinking around current capabilities, where we are, and what's coming. (Also, I have one "Intro to LLMs" video already from ~year ago, but that is just a re-recording of a random talk, so I wanted to loop around and do a lot more comprehensive version of this topic. They can still be combined, as the talk goes a lot deeper into other topics, e.g. LLM OS and LLM Security) Hope it's fun & useful! youtube.com/watch?v=7xTGNN…

Tighter regulation and a lack of high-quality, authentic training data are just some of the problems AI developers will be facing, according to @dswnds from @latrobe theconversation.com/ai-will-contin…

Happy birthday to Geoffrey Hinton, whose work in neural networks has led many to call him the "Godfather of AI." These pioneering efforts recently won him the Nobel Prize: bit.ly/3Bf92Vx Image v/Aaron Vincent Elkaim & @nytimes

This is the most important paper in a long time . It shows with strong evidence we are reaching the limits of quantization. The paper says this: the more tokens you train on, the more precision you need. This has broad implications for the entire field and the future of GPUs🧵

AlphaFold3 is now open source. Thanks @GoogleDeepMind @demishassabis openread.academy/en/paper/readi… github.com/google-deepmin…

"One must divide one's time between politics and equations. But our equations are much more important to me, because politics is for the present, while our equations are for eternity." Albert Einstein

Tools such as Rosetta and AlphaFold have redefined the protein-engineering landscape. But some problems remain out of reach — for now go.nature.com/3YT7zgN

Are prompt repositories the new frontier for #AI at work? Daswin De Silva (@dswnds), Professor of AI and Analytics at @latrobe, explores how a repository of #prompts can help organisations inform training programs, detect productivity gaps, and more. ceda.com.au/newsandresourc…

Hitchhiker's guide to rebranding: - Machine learning -> statistical mechanics - Loss function -> energy functional - Optimize the model -> minimize free energy - Trained model -> reached equilibrium distribution - KL divergence -> free energy difference - Gaussian noise -> random thermal fluctuations - Random step -> Brownian motion - SGD -> directional Brownian motion - GPU -> simulated particle accelerator - Diffusion models -> Langevin dynamics - Reinforcement learning -> control theory - Robotics -> physical computation - Audio learning -> 1D signal processing - Image learning -> 2D signal processing - Video learning -> 3D signal processing - Multimodal models -> multidimensional signal processing - Sora -> learned physics engine You're welcome

BREAKING: This year's #NobelPrize in Physics has been awarded to John J Hopfield and Geoffrey E Hinton for "foundational discoveries and inventions that enable machine learning with artificial neural networks" Stay tuned for more from @nature

Not fully sure why all the LLMs sound about the same - over-using lists, delving into “multifaceted” issues, over-offering to assist further, about same length responses, etc. Not something I had predicted at first because of many independent companies doing the finetuning.

With work at our fingertips at all hours of the day, it can be hard to disconnect. AI assistants may solve the problem – at least, that’s what big tech wants us to think, writes @dswnds (@latrobe). #Echobox=1708056820-1" target="_blank" rel="nofollow noopener">theconversation.com/drowning-in-di…