Francis Irving
banner
flourish.org
Francis Irving
@flourish.org
Coercing LLMs into understanding software product teams.
Cohost on "Imagine an apple" podcast about our inner experience.
Trustee of London College of Political Technology (soon).
He clearly really doesn't want to believe humans are machines, AND wants to believe physics can explain things. Which are contradictory desires, if you're studying minds and consciousness.
November 24, 2025 at 8:24 AM
Oh gosh, no-episodic-memory-me didn't remember you'd come to that too!

The thing I do remember (and you can maybe correct me) is I asked the question "But we could make computers that use the same quantum effects you say are used in our minds, so AI still possible right?", and he said "Yes"
November 24, 2025 at 8:22 AM
I found the book very emotionally connecting, making me warm and connected to the variety we have as humans.
October 5, 2025 at 4:15 PM
It's about how people think in lots of varying circumstances - from schizophrenia to playing the guitar. All based on a random mind-sampling technique, which is more like a therapeutic practice than a psychology questionnaire.
October 5, 2025 at 4:15 PM
Agreed, each metaphor is getting closer, more like a simile!

And I was having a long conversation with @berbank.bsky.social about the whole thing being compression just yesterday... Hadn't thought about the scientific method being that, but it is.
August 28, 2025 at 6:16 AM
We're doing that a bit with LLMs too. There's something very architecturally different about human minds that we still don't know/understand. LLMs help solidify things we knew anyway about the mind (it is made by composing trillions of connection weights), and give a feel for that. That's all.
August 28, 2025 at 5:42 AM
Julian Jaynes' The Origin of Consciousness in the Breakdown of the Bicameral Mind has a brilliant opening chapter explaining how we *always* make simile's of current technology to the human mind. e.g. It's like a steam engine.
August 28, 2025 at 5:42 AM
The various Anthropic interpretability papers a least give a hint at how it is doing it - with imagination of the few features they document scaled billions of times.

Yes, humans are similarish networks, and part of what we do is clearly "LLM-like". I wouldn't stretch this too far though.
August 28, 2025 at 5:42 AM
Reposted by Francis Irving
Just learned from @flourish.org’s “imagine an apple” podcast that permanent musical soundtracks are part of some people’s inner experience and here is another one!
January 31, 2025 at 4:54 AM
Those kinds of practical uses of imagination are easy and fun to talk about, and potentially help people learn ways to improve their lives by improving their minds.

zencastr.com/z/CCEuZGjW

Do get in touch if you use your mind in similar interesting ways!
Navigating a city with Anna by @Francis Irving
When you navigate a city, what is your inner experience? Do you see detailed overhead maps, or street-level views of landmarks, or neither? Vynn Suren and Francis Irving interview Anna about how she ...
zencastr.com
January 20, 2025 at 10:34 PM