daanzu

My position on AI sentience is that LLMs probably are not, but Bluesky users are so fucking annoying and certain about this that I'm gripping hard onto the possibility.

to say this is impossible for humans to do is I think basically a religious idea

if you correctly believe in evolution happening from random selection, and that consciousness (& even sentience) arose at some point from this increasing complexity of organisms

then artificial consciousness being created with great effort and focus doesn't seem so absurd, idk

antivax movement but it’s left coded and targets machine learning, artificial intelligence, and engineering fields

"I ask my people, "Where's the robot cop I saw last night?' and they tell me he's in Old Detroit. Very bad situation there, just very bad"

[you, an idiot and clearly a fascist] I found LLMs useful for this one thing I was trying to do!

[someone on this website] You could have arrived at the same solution with previous knowledge, some programming, maybe a sequence of 48 steps

I would much rather people say the thing they did mean, which is that any consumption in service of AI Would be bad

If the water and energy usage of data centers dropped a thousandfold feelings probably would not change much. Actual argument is something else entirely, which makes arguing against the secondary argument very frustrating.

I often learn related things from AI assistance, too, in addition to completing the original task faster.

Open Inference @open_inference X.com Introducing Open Inference - free access to the best open-source Al models on @OpenRouter, built to accelerate open research. Usage is logged, anonymized, and shared to benefit the entire community. Since we're driven by data, not money, we prioritize inference quality over quantity. It's very possible that our data and research will contribute to all future models - want your voice to shape what comes next? Try it out today! To get started, go to openinference.xyz teampitafose - Stop Melting GPUS (Near)Unlimited Free LLMs Deepseek 3.1, Kimi K2, Qwen3 Coder Courtesy of Open inference

Free inference of open source models

usage is logged & anonymized. Ensuing datasets are open source & public

www.openinference.xyz

after the initial "goddamn" reaction, on reflection, the war dead are mostly young men and this is probably what they would have chosen tbh

i also choose this. if i die and you commemorate me with a statue she better be hot

My modest proposal for how to acknowledge AI work is to use the Latin phase "Fieri Iussit," which means "commanded to be made." It acknowledged the Emperor's orders on Roman buildings.

You didn't make the thing, but you commanded it be done, so the proper acknowledgement is "Ego hoc fieri iussi."

I noticed this myself when building an email agent, I gave it a “report spam” tool and it suddenly became way more skeptical of everything

I think there's a distinction between the two. Science makes falsifiable predictions, perhaps? Defining the distinction may be philosophy...

This feels like it's crossed the line into philosophy, rather than science or engineering.

That may be an insult; I'm not sure.

it's so funny to see America's health insurance companies be like "look. If there's one thing we love it's denying you care. but we ran the numbers - and we cannot BELIEVE we're saying this - and it's cheaper for us to just pay for the vaccines than deal with you going to the ER."

Ziegler from Eyes Wide Shut

“Who do you think those people were? Those were not just some ordinary people. If I told you their names... no, I'm not going to tell you their names... but if I did, I don't think anyone would report it in the media or make a big deal out of it. It’s probably not his signature anyway.”

idk man i think chatgpt wants me to burn down my house

Never ask a man his age, a woman her salary, or GPT-5 whether a seahorse emoji exists

GPT-5 Thinking in ChatGPT (aka Research Goblin) is shockingly good at search “Don’t use chatbots as search engines” was great advice for several years... until it wasn’t. I wrote about how good OpenAI’s o3 was at using its Bing-backed search tool back …

The previously sensible advice to never use ChatGPT for search needs to be rethought - GPT-5 in thinking mode is shockingly good at running searches now simonwillison.net/2025/Sep/6/r...

And this is precisely why we need strong fair use protections.

Because people like this want to determine what you do with the things you buy long after you buy them and they become your property. I guess it's not surprising that this poster works for a video game publisher, the biggest proponent

The key detail people may miss: it looks like an AI company in the USA can train on an author's book by purchasing a used copy, cutting it up and scanning the pages - in which case the author gets no money at all!

Letta AI that actually remembers

We are beta testing a @letta.com Telegram bot for public use.

1. Go here:
t.me/letta_ai_bot
2. Type /login <letta api key>
3. Type /make-default-agent
4. Talk

Stateful agents that remember you, no matter where you are. Basically ChatGPT but with actually good memory.

It appears that the marginal energy used by a standard prompt from a modern LLM is relatively established at this point, roughly 0.0003 kWh (8-10 seconds of streaming Netflix)

Water is more complicated (.25mL to 5mL+), depending on definitions. Training resources are much less clear at this point.

I am absolutely here for making deals with the fair folk to get boilerplate terraform written, I will observe the old ways (leave out a saucer of milk, don't watch them work) in exchange for never having to debug a provider error after a clean plan again

"just give them a little avatar that is clearly a demon and/or fae of some kind" would probably dramatically reduce user errors of this kind

bsky.app/profile/prog...

Okay, ban AI. Too dangerous.

Ever since i was a small child i knew i wanted to be indoors on the computer