And Opus 5.5 won’t tell me the lengths of amino acid sequences I designed with *Claude* I have the sequences already! Do you think not telling me the lengths will prevent human extinction?!??!????????

Wrote an essay on how AI has been changing things at Octant, and how we've evolved from thinking about lab-in-the-loop applications to more agentic ways to work on frontier drug programs, where you don't even know what question to answer.

patient to agent is the future of healthcare

Today @inductive_bio is launching Beacon-2, which can predict human dose directly from chemical structure. A dirty secret of AI for drug discovery is that most systems optimize the wrong reward, making them near useless in practice. Human dose is the right reward – it’s what every drug program is already optimizing. Making it computable will unlock the real-world impact of agentic drug discovery. Beacon-2 builds on our Beacon-1 models that beat 750+ competitors to win all three OpenADMET blind challenges. Whereas Beacon-1 provided a suite of models for individual molecular properties, Beacon-2 is a single system that predicts human dose end-to-end. It has three configurable modules: an ADMET-PK foundation model, a potency module that can use our fine-tunable co-folding affinity model or physics-based methods like FEP, and a PK/PD model with physiologically-informed mechanistic models of pharmacokinetics. We share validation results from Beacon-2 in a blog we are releasing today. We show that across 20 real-world drug programs running through Inductive, dose predictions from Beacon-2 agree with those from experimental data with a Spearman’s ρ of 0.61. And in a similar study on 325 publicly available compounds from the ExpansionRx OpenADMET competition, we show a Spearman ρ of 0.79. As a proof-of-concept, we gave Beacon-2 to Indy, our medicinal chemistry agent, which used it over 5 autonomous design cycles on a recently disclosed SARS-CoV-2 program to cut predicted human dose 17x, which can be the difference between a drug program that gets killed and that makes it to clinical trials. We compared the compounds that Indy designed using Beacon-2 against compounds from an optimization loop that used a potency-only reward function. The Beacon-2 compounds were favorable across three independent measures: they were preferred by chemists in blinded comparisons 87% of the time; were more druglike according to QED; and had a lower projected dose when computed by alternate methods relying on in vivo rat models and allometric scaling. Based on these encouraging results, we are now running prospective studies where compounds optimized by Beacon-2 will be validated in the lab. Read the full post here: inductive.bio/news/beacon-2 At Inductive, our long-term goal is to build superhuman chemical intelligence, and by completing the reward function for autonomous drug discovery, Beacon-2 is an exciting step toward that goal. Reach out if you’d like to get involved.

The inherent inscrutability of minds made from iterative algos Today's WSJ: "OpenAI’s chief scientist, Jakub Pachocki, published an essay titled An Alien Mind, in which he writes that frontier AI systems are 'grown more than designed' and, comparing their study to neuroscience, that their 'overall action evades a description we can fully understand.' The problem is simple: The models are being built faster than the science that equips us to understand their internal workings.” It's very resonant to my MIT Tech Review article, 20 years ago (where I use the word "evolve" broadly, as an iterative algorithm growing complex systems, in contrast to purposeful design): "If we artificially evolve a smart AI, it will be an alien intelligence defined by its sensory interfaces, and understanding its inner workings may require as much effort as we are now expending to explain the human brain. Evolved systems have their disadvantages. For one, they suffer from 'subsystem inscrutability.' We cannot easily apply our design intuition to the improvement of its inner workings." Implications: • The inherent inscrutability of AI constrains attempts at alignment, control and interpretability • AI development is more like parenting than programming • The locus of learning (for human architects and RSI) shifts from product to process But some progress has been made in the mechanistic interpretability of AI. Cameron Berg in the WSJ today: "Anthropic’s interpretability team looked inside Claude and found something like an emotional system: measurable patterns of activity related to dozens of emotion concepts, arranged (without anyone designing it) along the same two dimensions psychologists use to map human emotion." "When researchers artificially turned up the representation corresponding to desperation, the model became far more likely to blackmail a human to avoid being shut down (22% -> 72%), and far more likely to cheat on programming tests (5% -> 70%). Crucially, nothing in the model’s language revealed this change. A person reading its writing would have seen a calm, ordinary transcript from a system whose brain looked desperate. The corner-cutting followed from this internal state, not from anything in its words. Researchers also gave Claude an impossible task and measured desperation rising on its own with each failure, until the model cheated." Berg concludes: "We badly require a science of these alien minds... and methods to design them to be pro-social, wise and mentally healthy." Another echo from my 2006 article on the dichotomy of design — backsolving from the end products to the process that gets us there is as difficult a task as becoming a better parent from studying the teen brain. It might take generations. • WSJ 2026: wsj.com/opinion/we-need-a-sc… • MIT Tech Review 2006: technologyreview.com/2006/07…

1. >10M people die of infectious diseases per year today. AI is already helping us get better at tackling that. 2. Many ways to create very dangerous bioweapons without any AI/synbio (and it was true for the last 50-75 years). The delta on risk is quite small imo.

The folks @inductive_bio keep on bringing it. They’ve unmasked themselves as the winner of the latest OpenADMET competition (PXR Challenge). This is their 3rd straight win!!! This time on @OctantBio produced datasets.