Feed

"The central challenge of brain emulation is not to store or compute the neurons and parameters, but to acquire the data necessary for setting neuron parameters correctly in the first place," he writes. ""I believe that to get to human brains, we first need to demonstrate mastery at the sub-million-neuron-brain level: most likely in zebrafish. For such organisms, like the fruit fly, a well-validated and accurate brain emulation model could be created in the next three to eight years… "Conditional on success with a sub-million-neuron brain emulation model, a reasonable order of magnitude estimate for the initial costs of the first convincing mouse brain emulation model is about one billion dollars in the 2030s and, eventually, tens of billions for the first human brain emulation model by the late 2040s."

I love the idea of having my website directly connected to the fediverse. A bluesky comment section is really cool. I instead turned my website into an ActivityPub actor. I don't have comments setup though. I bet I can do something similar.

There are other services that could be used for this purpose instead. Notably, I could embed replies from the social media formerly known as Twitter. Or I could use a platform like Disqus or even giscus, which hosts comments on GitHub Discussions. But I see Bluesky as a clearly superior choice among these options. For one, Bluesky is built on top of an open social media platform in AT Proto, meaning it can't easily be taken over by an authoritarian billionaire creep. Moreover, Bluesky is a full-fledged social media platform, which naturally makes it a better option for hosting a conversation than GitHub.

Projects, not jobs. Hiring, not being hired.

I had my clanker convert my personal website into an ActivityPub actor. It was a lot of fun when it actually did what I wanted it to. I also learned quite a bit about how ActivityPub works while trying to steer this thing in the proper direction.

Pi is written by Mario Zechner and unlike Peter, who aims for "sci-fi with a touch of madness," Mario is very grounded. Despite the differences in approach, both OpenClaw and Pi follow the same idea: LLMs are really good at writing and running code, so embrace this.

I think the upshot of all of this is that there is a lot of value to actively managing your agents' vibes and feelings, not just treating them as text-generation robots. The models aren't alive, but thinking of them as having feelings, rather than just next-token-prediction engines can help you nudge their mental states into a better place. I think you'll like the results.

As a business school professor, I think many people have the skills they need, or can learn them, in order to work with AI agents - they are management 101 skills. If you can explain what you need, give effective feedback, and design ways of evaluating work, you are going to be able to work with agents. In many ways, at least in your area of expertise, it is much easier than trying to design clever prompts to help you get work done, as it is more like working with people. At the same time, management has always assumed scarcity: you delegate because you can't do everything yourself, and because talent is limited and expensive. AI changes the equation. Now the "talent" is abundant and cheap. What's scarce is knowing what to ask for.

My bet is that most parts of cyberoffense and cyberdefense are going to move to running at "machine speed", where humans get taken out of most of the critical loops. This will both increase the frequency of hacking attacks while also dramatically scaling up the effectiveness of any individual human defender or attacker (as they will be scaled by AI systems which work for them). The true wildcard question is whether this turns out to be offense- or defense-dominant – my guess is we're heading for an era of offense-dominance as it'll take a while for defenses to get deployed.

With general intelligence, you can go a layer deeper: you can accelerate the acceleration. You don't just write the prompt that fixes the code; you build the evaluation pipeline that automatically optimizes the prompts. You stop working on the work, and start working on the optimization of the work. You shift from First-Order execution (doing the thing), to Second-Order automation (improving the system), to Third-Order meta-optimization (automating the improvement of the system). AI eats the lower derivatives, constantly pushing you up the stack to become the architect of the machine that builds the machine.

You can't leave anything on the table. This is Amdahl's Law for AI transformation: as the "core" work approaches zero duration, the "trivial" manual steps you ignored—the 10-minute deploy, the manual data entry on a UI, the waiting for CI—become the entire bottleneck. The speed of your system is no longer determined by how fast you code, but by the one thing you didn't automate5. If an agent can fix a bug in 5 minutes but it takes 3 days for Security to review the text or 2 days for Design to approve the padding, the organization has become the bug. You need to treat organizational latency with the same severity you treat server latency.

I'm still skeptical of vibecoding in general. As I mentioned above, I would not trust my vibecoding enough to make these into products. If something goes wrong, I don't know how to fix it. Maybe my LLM friends can, but I don't know. But vibecoding is 100% viable for personal stuff like this: we now have apps on demand.