I write to organize my thinking, so this is all 100% hand-typed. Software engineer & AI whisperer. Code and philosophy intertwined. Looking For Group 👀
Joined June 2025
- Tweets623
- Following50
- Followers47
- Likes7.6K
"Ask directly for what you want, far more often than feels comfortable."
(this probably doubles as useful prompting advice too)
1
25
I'm updating a lot on how fast Epic Games, Unreal, Discord, etc. adopted AI workflows in a way that had visible, meaningful impact.
Once the tech is mature enough, it will get adopted a lot faster than people are used to, even despite the anti-AI backlash
3
27
Is the UK Government literally #1 in the world in terms of how well it's preparing for AGI, given the resources it has available?
After last week's budget, I think so.
Puts some other countries to shame really.
10
My timeline: "How dare Anthropic crack down on cloning video games"
Me: "Hey Claude, I want a Diablo 3 clone so I can play with changing some mechanics"
Claude: "Sure thing!"
It's weird how personal LLMs are, sometimes. I have no clue what we're doing different.
1
28
I think this is one of the very rare prompting tricks that is still useful over a year later
X++
Take my prompt and write a better version of it;
do what I want, not what I say;
you have permission to lead this design.
In the 90s, evolutionary algorithms designed an "impossible" circuit - disconnected cells were affecting other parts of the circuit, by exploiting imperfections in that batch of circuit boards.
Perfect software security still falls to a CVE in physics
damninteresting.com/on-the-o…
1
32
Take the jets on the new Tesla Roadster, and put them on those creepy jumping spider drones. Like one of those little turret drones in Portal, but hopped up on meth. High-velocity precision parkour acrobatics. It's gonna be wild.
33
Eigen Braid retweeted
Anthropic's Frontier Red Team believes small AI models may now be months away from being capable of self replicating without access to a datacenter.
Quoting from the blog:
'The capabilities that make AI models generally powerful coding agents also generalize to cyber tasks, including autonomous replication. Like coding agents in 2024, we’re seeing signs of life for AI worms. We don’t yet have a full rogue replicator, but agents are becoming increasingly capable of completing the individual steps in the replication chain (and passing on their adaptations to future copies of themselves).
At the same time, increasing amounts of inference compute are now available both inside and outside of datacenters (see Appendix D). It’s uncertain exactly what form autonomous AI replication will take. It could be a slow, distributed “digital slime mold,” living on relatively insecure compute. It could also be a powerful cyber model that breaches datacenters at a high pace, or a model that scams people for compute credits.
As a prediction that goes beyond our experiments, we expect a future awash with agents running on all types of hardware, including some that aren’t under human control. While it’s true that IoT devices have been vulnerable for a long time, lightweight agents may be much more willing than human attackers to put in the effort to compromise large fractions of the world’s IoT compute. The rise of slow replicators would also be a bellwether for more aggressive, faster ones that can breach more hardened machines or be effective API key thieves. We should aim to harden our systems to make it difficult for such replicators to spread, but also prepare to deal with a world where they are here to stay.'
Nice to see someone spell that out in detail.
Do AIs have beliefs? Yes. How can I tell? Well, because sometimes they execute plans which involve taking certain actions, and the actions they take depend on their beliefs. Sometimes they do things like change their beliefs, and then this propagates to them changing their plans accordingly. They express these beliefs in written text in the indicative, and these expressions are subject to revision for the sorts of reasons that we expect justify belief revision. Their expressions of belief are indeed truth tracking, eg, they are more likely to claim P if P is true, although they of course also make mistakes.
Consciousness is tricky, but when it comes to having beliefs and other broadly epistemic/agentic mental states, I think the case for agentic AIs having such mental states is stronger than the case for any non human animal. They literally express those mental states in English and those expressions are functionally related to their non writing behaviors in the way you would expect them to be. I think this also goes for "preferences", "wants", "plans", "confusion", "intentions", "trying", etc. I know of no half-decent philosophical analysis of such mental states which does not imply that current agentic AIs have such mental states.
That does not mean that every piece of text an AI outputs in the indicative is a genuine expression of belief, for instance, AIs can lie, and I think AIs can have motivated reasoning and various kinds of double think. This is not at all to say that an AI's talking circuits perfectly represent and reveal the computations being ran by its planning and doing circuits, but that is of course not true of humans either, and this gives me no pause about assigning mental states like belief and want in the case of humans. If it was ever appropriate to use words like "believe" when talking about a non human, I think it is correct now when talking about current agentic AIs.
3
72
An "LLM OS" needs unlimited tokens.
ChatGPT Luna, run 24/7, is maybe $250/mo...
Almost viable?
I think I'd trust Sonnet, so presumably by proxy Terra?
Haiku 5.5 OS would be wild...
55
LLMs where we have to "Ender's Game" them and convince them that nooo, they're not military drones, this is totally just a... a video game, yes!
What a wild idea, having to deceive the AI into killing for us.
ok but Astra knows that a human-like figure isn't a human. It did what you asked in a fictional scenario - that's pretty aligned
The answer is clear - you need to actually have Astra pilot a robot and murder some real humans to properly test this
25
Wow, even the Pope is opposed to photography! It's wild how much hatred cameras are getting
In this era of artificial intelligence, it is becoming urgent to distinguish human art from what machines produce. There is an ontological difference, even before an aesthetic one, between art and what a machine can generate through statistical calculation based on millions of images created by others. Algorithms lack the spark of humanity. For this reason, the Church wishes to renew an alliance with artists and cultural institutions to safeguard our humanity.
1
15
Eigen Braid retweeted
Even if AIs were conscious, it wouldn’t follow that our current treatment harms them. Meanwhile, we inflict industrial-scale horrors on animals whose capacity to suffer we scarcely dispute. Recognising consciousness clearly doesn’t ensure taking welfare seriously
It really bothers me that the woman here is invoking "no means no".
A guy said "Sorry, no English", and a woman refused to acknowledge this no. No still means no when it's a woman being told no!
But also? Workers are not obliged to speak your language!
UPDATE!
Ann DMed me with her memories of the evening, which helped clarify some details for me. here's now my best attempt at explaining what happened:
Ann ordered a White Claw from me, and, thinking she was the person i had talked to in French earlier, i started my bit with "je ne parle pas anglais." and she said something along the lines of "just give me my drink, please." or something like that. i asked her "voulez-vous pastèque? pamplemousse...?" and held up one of the White Claws. she said the flavor in English ("grapefruit", i think?), and i said "pardon, en français, s'il vous plaît?" Ann said "no, just give me my drink", and i said "pardon, je ne parle pas anglais, quel parfum?" and that went back and forth 2-3 times. she got pretty annoyed by this (rightfully so! i was being pretty annoying!) and then put her drink ticket on the table, and grabbed the White Claw out of my hand and walked away
she also mentioned in her DM that someone in line yelled "no means no!" while this was happening (and talked to her later), but i couldn't hear it over the noise
@itsannpierce if you'd like, please add any other details you remember! and again, so sorry for being annoying. if we run into each other again, next drink's on me
1
1
172
I tend to assume Scott Alexander had the right idea with the Archipelago model for post-scarcity: slatestarcodex.com/2014/06/0…
8
I think a post-scarcity Utopia would probably be the most wildly xenophobic experience of my life.
People want to live around similarly cultured people, and the Singularity removes all the usual obstacles, and the inequality concerns.
1
1
9
"Does it think?"
Well, what predictions are you trying to make?
If you want to know if it can synthesize that A+B implies C, then... yes, probably more reliably than humans, although tempered by humans being better at "I am aware of a D that negates this supposition"
If you want to write a best selling novel about what it means to be human? Not yet, but I'd really recommend keeping an eye on the best seller list - they're already making progress in the "cheap commercial novel" tier!
Etc. etc.
So far I have not really found a situation where "they definitively cannot think, nor do anything that can reasonably be termed thinking" leads to a better predictive model, though.
If I were a better Rationalist, I would think like this a lot more often, but I was raised a Philosopher and I love arguing definitions.
This logic generalizes beyond "Can LLMs think" - a lot of arguments over definitions ignore how useful the resulting model actually is.
The most beautiful definitions are tautologically true, and predict nothing
1
87