Roly Perera

Excited to be speaking on Subjectivity as Self-Simulation: Virtualising the Cartesian Theatre at the AGI 2025 Philosophy and Machine Consciousness workshop. Even more buzzed to see @annaciaunica.bsky.social and @drmichaellevin.bsky.social just announced as invited speakers! agi-conf.org/2025/ 🤯

Workshop on Live Programming (LIVE) The 11th Workshop on Live Programming (LIVE 2025) will take place online. LIVE invites submissions of ideas for improving the immediacy, usability, and learnability of programming.

Are you working to make programming more direct, tangible, and alive? Think about submitting to the LIVE 2025 workshop — Call for Submissions open now: liveprog.org

Cannot make it to Prague for @programming-conf.bsky.social?

🎬 We will be streaming the keynotes, regular talks and MoreVMs workshop talks online!

👉 Keep an eye on: 2025.programming-conference.org

Introducing The Darwin Gödel Machine

sakana.ai/dgm

The Darwin Gödel Machine is a self-improving agent that can modify its own code. Inspired by evolution, we maintain an expanding lineage of agent variants, allowing for open-ended exploration of the vast design space of such self-improving agents.

@zoedrayson.bsky.social I read (and enjoyed!) this article recently - I’ve been interested in subpersonally-held beliefs with personal-level content. In a sense a lot of our self-conception falls into this category but was wondering if there was much discussion in the philosophy of mind literature?

Sounds interesting! But what are these mysterious “physical events constituting reality” of which this author speaks?

These boxes are not moving. A mind-bending optical illusion by Japanese artist Jagarikin.

I remember having that feeling about the Moon when i was about 4 years old, when i first realised that sliver in the sky was a big ball backlit by the sun.

What’s wrong about IIT is closely related to what’s wrong about the Hard Problem. There is no Hard Problem of consciousness that science needs to solve. Science only needs to explain why it seems like there’s a Hard Problem. But that’s a non-Hard problem, by definition.

Old man goes to church a bit, and dies. Possibly of old age.

The basic problem with IIT is that it takes our intuitions about consciousness and treats the content of those intuitions as the observable to be explained. But the actual observable to be explained is simply the fact that we have intuitions with that content. Explaining that would cover all bases.

(All organisms are in the business of self-evidencing, but with consciousness the evidence-gathering actions — introspection, planning, inner speech, etc — seem to be characteristically “covert”, taking place not in outwardly observable physical space but in some kind of internal epistemic space.)

We interpret ourselves as inhabited by immaterial souls. That’s a good folk theory because we act *as though* inhabited by immaterial souls. But that self-interpretation is the very mechanism by which we act that way in the first place. A self-justifying theory, bootstrapped over evolutionary time.

I think of consciousness as a kind of tacit knowledge: knowledge of what it would be like to be you if you were conscious. It turns out that's all you need to actually *be* conscious. (Fake it till you make it!) Unpacking a bit:

I love that a human artifact is a full light-day away

Pickle pizza is now a thing — and it took years of walking around NYC to develop It’s kind of a big dill.

The internet age is characterized by extremism of all kinds at all times, including highly secular, cultural extremism.

One cannot simply enjoy, say, pickles anymore. Pickles must become an obsession, a first principle, an entire worldview.

One of the biggest contributions of FEP is that it shows us what a metaphysically neutral “physics of sentience” looks like: it shows how you can always *read* systems that maintain themselves far from thermodynamic equilibrium as modelling their environment and acting to maximise model evidence.

I think the formal framework (or axiomatic) nature of the FEP means that FEP itself isn’t directly falsifiable. But specific deployments of the FEP on concrete problems are. IIT is at the other extreme: it starts with a strong (and unnecessary) metaphysical claim and then makes that the explanandum.

I’m also skeptical of frameworks with metaphysical baggage. I’d put IIT and FEP at opposite ends of the baggage spectrum, though. Not all proponents of FEP may be as careful but Friston’s position is quite deflationary: systems “can be read as” having self-models, doing Bayesian belief update, etc.

👀 Fancy a summer internship working at the intersection of programming languages, AI/ML and HCI, working on a new language being developed at ICCS for open climate science communication? See @f.luid.org and www.epsrc.group.cam.ac.uk/DTP/internsh... for more information.

⏰ Deadline Fri 14 March!

More pithily: the explanandum is why we take ourselves to be the way we take ourselves to be, not why we *are* that way. Answering that explains everything there is to explain, subjectively and objectively.

I’d go further and say that it’s not really scientific to even recognise the existence of the so-called Hard Problem. The question of why we think there’s a Hard Problem is a scientific one. The Hard Problem itself is not.

I’d have 100% been up for this. Let’s hope they make it to SIGGRAPH (or something similar) in 2025.

I love idea that if you wait around for a while there are enough photons for this

For example, it becomes possible for a web extension to turn any web page into a spatial canvas and start deconstructing it!

Amazing Cyrus, congratulations!

If you can have dragon fruit then you can have coconut meat

This sounds great. So much of understanding subjectivity/consciousness as natural phenomena seems to be tied up with the role of covert action and epistemic agency.

“Patience is also a form of action.” — Auguste Rodin

The only tweet I’m salvaging. Bye Twitter.