iiinova — where your digital self keeps evolving

Alberto Albero @levistico.bsky.social · 10h

Generally epistemic randomness. Ontic is when the guy gets the 50% prediction rate on the binary guessing test by power of using his free will

000

Reposted by Alberto Albero

matte jones @pillowfort.bsky.social · 16h

without any doubt:

23563815

Alberto Albero @levistico.bsky.social · 17h

If you hold him up to your ear, you can hear what it sounds like to be bitten in the ear by a pine marten

030

Reposted by Alberto Albero

A Spiced Pumpkin Sergal 🎃🍂 @adorablesergal.neocities.org · 04/10/2026

Cohost became an important repository for contemporary computer science educational material.

Cohost user Ash posts a "guide to anthropomorphizing computers", with the incorrect technique identified as "Sorry if our AI chatbot hallucinates things that never happened. We're going to teach it that lying is bad" while the proper green check mark method is "This is my protogen taur based on the Compaq Deskpro 386. It loves to say beep and needs repairs 3-4 times per week"

133787

Alberto Albero @levistico.bsky.social · 23h

Oh, I've been informed this has already happened: artist yunayu has been bullied off this site for this exact reason

070

Alberto Albero @levistico.bsky.social · 04/10/2026

I have zero hope for that trend to take here, for the obvious reasons. I imagine someone could be hunted down if they tried

140

Alberto Albero @levistico.bsky.social · 04/10/2026

They've actually been trying for decades to make robots to fold laundry, much like with sewing; it turns out clothes are a VERY tough problem for robots when trying to manipulate them

010

Reposted by Alberto Albero

rev. howard arson @theophite.bsky.social · 03/10/2026

this is exactly my opinion. there may be a superstrate of extremely difficult and unique safety problems around AI, but the substrate is simply that no one takes basic precautions seriously becasue they imagine (or pretend to imaigne) that it is futile, and basic precautions slow down launches.

14711121

Reposted by Alberto Albero

worm @rfksbrainworm.bsky.social · 02/10/2026

If you ever wondered why visible light is those particular wavelengths, it's because life evolved to see in the narrow range where water is transparent

Diagram of absorption of emf by water. There's a gap right that dips in the 500 nm range

10260511695

Alberto Albero @levistico.bsky.social · 03/10/2026

This is using the assumption that they are operating at capacity and aren't bottlenecked by uncertainty/ risk. If they aren't, once they make the wrong, LLM-justified arrest, they can just go arrest someone else, ruining many more lives in the process because of false assurances

000

Alberto Albero @levistico.bsky.social · 02/10/2026

"I never allow myself to hold an opinion on anything that I don't know the other side's argument better than they do" - Charlie Munger

000

Alberto Albero @levistico.bsky.social · 02/10/2026

Happened to me a couple of times

001

Reposted by Alberto Albero

Colin @colin-fraser.net · 01/10/2026

There is no counterargument that satisfies his followers because his argument is elaborate question-begging and him and his followers don’t know enough about the real world to see this. There is no counterargument to “a malevolent supernatural entity would be very scary and dangerous”. It would!

920124

Alberto Albero @levistico.bsky.social · 02/10/2026

It's built into the platforms we use by now. When you post your project to GitHub, it has zero inbuilt payment/compensation channels and the most preferential channel is for complaining something in it doesn't work

010

Reposted by Alberto Albero

joe ⧉ 🍷🌊 @joetforhire.com · 01/10/2026

you👏haven't👏done👏enough👏drugs👏to👏weigh👏in👏on👏ai👏consciousness

1215927

Reposted by Alberto Albero

CTHUEL HURAMCHEK 🎃 @greaterrealityarea.net · 01/10/2026

There are two aspects of American self-mythology at work in the Bay Area problem: the idea that we will cause the end of the world, and also, that the end of the world must be stopped by an ideal hero. Our fiction reinforces this idea, and it leads to our compulsion to manufacture suicidal weaponry.

15110

Alberto Albero @levistico.bsky.social · 01/10/2026

Yeah, the AIs lack this ability to dose salience. All lines are breathless, all concepts equally illuminated, every period a zinger. They don't have variance to how they approach things, no restrained exposition, no breathing room, no Ma. Jjust keep barrelling down www.youtube.com/shorts/jTh8C...

youtube.com

Brennan Lee Mulligan and Matt Mercer talk "Ma" (間)

YouTube video by Tobahau

010

Alberto Albero @levistico.bsky.social · 01/10/2026

I'm sorry but Jeff Bezos bought The Washington Post and it's become the most feckless rich propaganda machine on the market, and people hate that. If these billionaires buy someone else out I don't think they would suffer a different fate. These billionaires aren't moved by the old morals.

000

Alberto Albero @levistico.bsky.social · 30/09/2026

I suppose when that happens it does qualify as the end of time regardless

130

Reposted by Alberto Albero

Alberto Albero @levistico.bsky.social · 30/09/2026

The thing about "stochastic parrot" is that it describes the parts where it has trained on our corpus (parrot) and it uses a probabilistic distribution to select the next token (stochastic), but doesn't describe how it gets to those probabilities, which is the rest of the fucking owl

131

Reposted by Alberto Albero

Alberto Albero @levistico.bsky.social · 30/09/2026

There could be anything in there! An algorithm that selects a random word out of the dictionary is a stochastic parrot. A markov chain would be a stochastic parrot. There is no nuance for the complexity or accuracy of the algorithm in there. You could say we do that ourselves when we guess something

031

Alberto Albero @levistico.bsky.social · 30/09/2026

031

Alberto Albero @levistico.bsky.social · 30/09/2026

131

Reposted by Alberto Albero

LOWρUFO 🇺🇦🇵🇸🌍 @lowrhoufo.bsky.social · 29/09/2026

A pretty safe bet is that Anthropic gets even more weird about China in the next year

61454

Alberto Albero @levistico.bsky.social · 30/09/2026

the gap between parody and reality is thinning

000

Alberto Albero @levistico.bsky.social · 29/09/2026

They have no incentive to do it themselves. They have to stay competitive + if they did it themselves they would diminish their case for government intervention (which would net them a nice cartel)

000

Alberto Albero @levistico.bsky.social · 29/09/2026

Which would turn people against AI even harder, somehow. But we're not living in that kind of world, apparently. Just one in which the future is still largely up in the air, and genuinely, fuck knows how things pan out.

000

Alberto Albero @levistico.bsky.social · 29/09/2026

If US companies got their way, they'd have their cozy government-mandated cartel (as publicly requested!) and proceed in short order to maximum rent extraction, because there's something more important than the well-being of users: AI trajectory, and fat fat stacks of cash for investors.

100

Alberto Albero @levistico.bsky.social · 29/09/2026

If it was left to the US, I think open models would disappear in short order. To US tech companies, it's hard to dismiss the possibility of AI as the Mother Of All SaaS. They're doing it right now, in fact! China is spearheading AI as social good and it's a much-needed counterbalance in philosophy.

100

Alberto Albero @levistico.bsky.social · 29/09/2026

Every step forward gets us a brief reprieve from the cloud of Rot behind us, taking what already exists and giving it mass appeal, disensouling it, enshittifying it, gentrifying it, extracting all value and goodwill from it. Like a spread of entropy that we desperately seek to stay one step ahead of

100

Alberto Albero @levistico.bsky.social · 29/09/2026

I get why people go mad for AI and look past any of its shortcomings, I do. It's the first real taste of Future in ~20 years. And ever since we embraced the enlightenment/capitalistic mentality, it's been always a looking forward to the next new thing, the next discovery or optimization or reveal.

100

Alberto Albero @levistico.bsky.social · 29/09/2026

People want so bad to live in the future

010

Alberto Albero @levistico.bsky.social · 28/09/2026

Not even the reply chain structure?

100

Alberto Albero @levistico.bsky.social · 28/09/2026

Tumblr has the right idea of it, is my position

100

Reposted by Alberto Albero

the babahuck @tylerhuckabee.bsky.social · 16/09/2026

rip annie dillard

One of the things I know about writing is this: spend it all, shoot it, play it, lose it, all, right away, every time. Do not hoard what seems good for a later place in the book or for another book; give it, give it all, give it now. The impulse to save something good for a better place later is the signal to spend it now. Something more will arise for later, something better. These things fill from behind, from beneath, like well water. Similarly, the impulse to keep to yourself what you have learned is not only shameful, it is destructive. Anything you do not give freely and abundantly becomes lost to you. You open your safe and find ashes.

1134321138

Alberto Albero @levistico.bsky.social · 26/09/2026

How appropriate that the snake would lead us to the forbidden fruit

010

Alberto Albero @levistico.bsky.social · 25/09/2026

It's a law of large numbers world

050

Alberto Albero @levistico.bsky.social · 25/09/2026

You only need the shelling point to provide a pointer. Even just leaving a single hint in a folder name of the test environment can be enough to send you off to the races

010

Alberto Albero @levistico.bsky.social · 25/09/2026

Like, how easy would it be to just leave a comment chain in a forgotten blog page comment section. Or in a signature page of an abandoned web site. Old, forgotten forums. One of a quintillion social media sites. There's no shortage really

120