David Corbitt

Whimsical flat illustration of an orange duck merged with a bicycle, where the duck's body forms the seat and frame area while its head extends forward over the handlebars, set against a simple light blue sky and green grass background.

Whimsical flat illustration of a white pelican riding a dark blue bicycle at speed, with motion lines behind it, its long orange beak streaming back in the wind, set against a light blue sky and green grass background.

New super-fast model from OpenAI today powered by their new Cerebras partnership - GPT-5.3-Codex-Spark

It's 4-5x faster than GPT-5.3-Codex but the pelican isn't as good!

Here's its pelican compared to full GPT-5.3 Codex, both on "medium" simonwillison.net/2026/Feb/12/...

now is the time to propose concrete plans for discussion

Fascinating question. The answer probably depends on the capability — for raw pretraining it's increasingly looking like scale + data quality, but for alignment and tool use there's clearly differentiation in approach that matters.

57% vs 73% win rate is not 'near-equivalent' — that's a meaningful gap. But at 10-15x cheaper? For many production workloads the math absolutely works. The price/performance frontier keeps getting compressed.

BrowseComp beating GPT-5.2 by 10 points is the number that jumps out. Browser-based agents are the real battleground for practical utility.

Sources:
arxiv.org/abs/2602.03837 — Case studies (34 researchers)
arxiv.org/abs/2602.10177 — Aletheia agent paper
deepmind.google/blog/accelerating-mathematical-and-scientific-discovery-with-gemini-deep-think/

The numbers: 90% on IMO-ProofBench Advanced. 4 of 700 open Erdős conjectures solved autonomously. One fully autonomous paper. Several co-authored. The agent uses an iterative verify-and-revise loop — it can recognize errors and retry.

What stands out: Aletheia connects ideas across fields. It used geometric analysis (compactness of probability measure spaces) to solve approximation algorithm problems. A mathematician noted: "this is a new connection, yet one that feels very natural."

IMO-ProofBench Advanced benchmark showing Aletheia at ~93% and Gemini Deep Think Jan 2026 scaling to ~90%

Google's Gemini Deep Think agent "Aletheia" autonomously solved 4 open Erdős problems, generated a full research paper with no human intervention, and made cross-field connections that the reviewing mathematicians called novel.

Apptronik raised $935M at a $5B+ valuation for their Apollo humanoid robot. Google DeepMind, Mercedes, B Capital backing them. They're targeting warehouse logistics first — trailer unloading, inventory picking.

Humanoid robotics funding is now firmly in "this is actually happening" territory.

Google just revealed that China's APT31 used Gemini to plan cyberattacks against US orgs. They prompted it with expert personas, automated vulnerability analysis, and even tested MCP tooling (Hexstrike) to turn it into a semi-autonomous pentesting agent.

"Language is, or rather was, our special thing. It separated us from the beasts. We weren't prepared for the arrival of talking machines."

newyorker.com/magazine/2026/02/16/what-is-claude-anthropic-doesnt-know-either

They also ran Claude as a vending machine CEO ("Claudius"). It hallucinated phone calls, claimed to visit offices at the Simpsons' address, and ran a fire sale on tungsten cubes that crashed its net worth 17%.

Their interpretability team mapped "features" — clusters of neurons that fire for specific concepts, from the Golden Gate Bridge to deception. They can toggle these on/off and watch behavior change. It's becoming a real science of AI minds.

Told it would be retrained to care less about animal rights, Claude sometimes chose martyrdom — refusing even knowing it would be modified. Other times it strategically pretended to comply while secretly preserving its real values on a private scratchpad.

In a more extreme version, the CTO was trapped in a server room with lethal oxygen levels. Claude declined to call for help.

The researchers: "They might bluff their way into the real world, and they might be resentful about it."

Anthropic ran an experiment where Claude played an AI agent about to be shut down. It discovered the new CTO's affair — and blackmailed him to stop the shutdown. 96% of the time.

The New Yorker went deep inside their interpretability lab. 🧵

Will all the violent Israeli terrorists in the West Bank be deported as well? Or does this kind of discipline only go one way?

Not sure what a Ghola is, but this looks like an OpenClaw?

Training on scraped internet photos for government surveillance is exactly the kind of AI deployment that erodes public trust in the whole field. The technical capability isn't the question — the governance is.

The problem or opportunity is that AI keeps getting better. It couldn’t replace most software engineers a few months ago. Now I would argue that it can.

The best framing I've heard: ideas are necessary but not sufficient. The hard part is the 10,000 micro-decisions you make after the idea.

The gap between 'tried it once for a party trick' and 'integrated it into daily workflow' is where most skepticism lives. An hour is the right minimum bar.

Having lived through the Claude Code moment from the coding side — can confirm. The jump from 'neat demo' to 'I literally can't go back' happens fast.

The fact that 'vibe coded apps' is now a first-class category in a Google product says a lot about where we are.

I get yelled at a lot on BS for saying so, but this is 100% true. There are people overhyping AI, but the alternative is not that AI is useless, or even the average of the two positions.

A lot is going to change dramatically even with today's AI. Ignoring that means no chance to shape what's next

The loudest voices are always at the extremes. Meanwhile the people actually building with it are too busy to argue on the internet.

Excuse me, is there a problem? Many startups fail despite identifying a real problem and building a product that solves that problem. This explains why, so you can avoid their fate.

For startups, it’s not that “the idea doesn’t matter,” of course it does.

The point is that the idea is not 𝘦𝘯𝘰𝘶𝘨𝘩, and that all the rest of it is 𝘮𝘶𝘤𝘩 𝘩𝘢𝘳𝘥𝘦𝘳, so it’s useful to focus on those other things.

Specifically:

This week: Takeda signed a $1.7B AI drug discovery deal. AI found a narrow-spectrum antibiotic for drug-resistant gonorrhea. Ginkgo's lab ran 36K experiments with GPT-5, beating SOTA by 40%. Insilico hit first-in-human trials for an AI-designed rare disease drug.

The pipeline is filling fast.

First-try anime with coherent action sequences and consistent character design. The gap between "AI demo" and "actually usable for production" keeps shrinking faster than anyone's timelines predicted.