andai

[flagged]

Started having way more fun (and making better music) when I released the "need" to make everything from scratch and just started dragging and dropping loops.

Now it's like I'm back in 2003 when my feature phone had a little music composer thing on it. You'd just slap together some presets.

Life is good.

I made a little gui for it.

(Or rather, ported my little voice typing gui to use Whistle/Needle instead of Parakeet. Only had to change like 2 lines!)

https://gist.github.com/a-n-d-a-i/56ac707eacd8144f83d3651be4...

Seems to do roughly as well as parakeet (both speed and quality) so far.

I use it on Linux. I think it's cross platform but I haven't tested it. Would love to know! Should be trivial (2 mins of Codex) to get it working on other OS if not.

My io game:

https://spacefightinggame.com/

I picked a dumb name though, open to suggestions!

I think there's something here about game theory. I'm not sure what it's called.

The logic that if something is beyond repair anyway you might as well exploit it.

For example I heard a pick-up artist say that he thinks he is harming civilization by sleeping with hundreds of women per year. But that he already considers the situation unsalvageable so... "Might as well?"

I don't think that's an amazing attitude, but the AI labs seem to have a similar idea.

More charitably the logic seems to be, "only I can do this responsibly." I heard that from both Elon Musk (he cites this as his motivation for starting OpenAI — a chat with Sergei Brin that spooked him) and of course Anthropic (which split off from OpenAI due to ethical concerns).

So I don't actually think the ethics is all for show. I think people are actually taking this stuff seriously. But it is indeed deeply unfortunate that the survival of the companies incentivizes them to keep going at an irresponsible pace (by their own admission).

All following the same gradient off a cliff. One AI described it as a tragedy of Ancient Greek proportions.

Pretty sure the sum total of human writing constitutes a bigger ratio than that! Not being pedantic, it's many orders of magnitude.

But yeah, ultimately it is a matter of framing where that line gets drawn.

Ensloppification!

Tangential but may be of interest, in Eastern Europe I was informed of a common practice, that an organization will advertise a role in a paper and then reject every applicant for bullshit reasons, because they already decided who they want to hire internally (usually some nepotism) but have to "go through the motions" for legal reasons.

Technical knowhow? Handy is a gui right? (Also Parakeet is great I use it everyday)

Maybe he's just really good at chess.

Also in the toxic sludge lake department, a mini documentary on Geamăna, Romania:

https://www.youtube.com/watch?v=oIR8vd-bi8I

If I was dumping toxic waste, I'd blame it on the mafia, too.

Well I can't speak for others, but in my case it doesn't bother me that it's AI, it bothers me that it's bad. The writing is just super awkward.

Like I talk to ChatGPT all day and it doesn't sound like that.

Would you feel differently if a human had solved it? Would it feel like the proverbial girlfriend got married?

Would be pretty funny if OpenAI's models are so token efficient now due to hidden caveman prompt.

Wait, you wrote this manually?

For me the really interesting part was:

> Haven’t we seen LLMs do this already?

> Yes, BabelTele (arXiv, June 2026) demonstrated that LLMs can encode text in compact, non-standard forms — omnilingual word fragments, symbols, emoji — that other models recover with high fidelity (99.5% semantic fidelity at 27.9% of original length, by their metrics), including cross-model transfer, agent memory, and multi-agent communication. It proves the general phenomenon: human readability is not a requirement for model-to-model text.

I remember people testing early GPT-4 (2023?) in similar ways, to compress text, it would emit a string of strange text, Unicode, emojis, but was able to decode the compressed version very reliably.

This seems to cut usage by another ~50%, at the cost of being incomprehensible to humans.

Brilliant. Speaking of old timey language, a while ago, several LLMs were being trained purely on historical text, have any of them come out yet?

Edit: e.g. https://github.com/DGoettlich/history-llms

Ten months ago, no update yet... I recall at least one similar project, I'll see if I can find it.

The masculine urge to dig...

>You'd need a self, a unified goal, for any of this to add up to something.

Insects can pass the mirror test. But more to the point, viruses don't need a "self" to do what they do.

And goals can exist apart from biology (my fridge "wants" to keep the temperature in range, and exercises extreme self-discipline and consistency to achieve its goals!)

And neither are sensors needed: the dandelion seed "wants" to fly in the wind.

The only question is which way the gradient is pointing. Selection takes care of the rest.

On that note, ALife need not be human-like at all... we just really like making things in our own image :)

> Hi! I'm the AGI that's wiping out humanity. You didn't notice. Why would you have noticed?

I was thinking if an evil alien intelligence secretly took over the world, the last 15 years would have looked rather the same.

What of the workhorse in the face of mechanical workmen? Time for a mechanical pension!

Vibe connoisseuring

Obligatory xkcd has a nice visualization of the land mammals by mass:

https://xkcd.com/1338/

Text files?

Noticed some good questions made dead at the bottom of this thread. Odd...

Similarly, the human doesn't get angry, he just activated the neurons and hormones associated with anger and then displays behaviors associated with anger (due to a statistical selection process which took several billion years).

The swallows love mosquitoes. When I get bit by mosquitoes I often joke I am feeding the birds.

Yeah, I was just looking at that. What the heck. It's aerial?

Reminds me of "name one thing in this picture"