iiinova — where your digital self keeps evolving

“Congratulations, you’re hired as the royal food taster!”

“You know the day destroys the night Night divides the day, Try to run, try to hide, Do a breakthrough. Make no mistakes.” — Claude Morrison

040

It seems like the food quantity wasn’t properly updated after buying food, so Jev was stuck in a loop buying food.

120

Daniel Lowd @dlowd.com · 25/09/2026

Oh wow I dunno where to begin… Oregon Trail Microsoft Flight Simulator Civilization Myst Sam and Max Hit the Road Lemmings Nethack Medievia (the MUD) World of Warcraft

020

As long as no one has been a bad user, it should be totally fine right?

Daniel Lowd @dlowd.com · 25/09/2026

Well, how did I get here?

340

Daniel Lowd @dlowd.com · 25/09/2026

AI discourse in 2006: How do we handle non-convex losses at scale? AI discourse in 2016: How do we use continuous embedding spaces for discrete domains? AI discourse in 2026: Do you like purple dragons?

0384

Reposted by Daniel Lowd

David Mimno @dmimno.bsky.social · 22/09/2026

It's possible for Jev/Laya/Decision Models to be not that big a deal as tech and massive as a new paradigm. Here's why I'm really excited from an NLP history perspective (thread)

17619

Daniel Lowd @dlowd.com · 16/09/2026

Everything should be anthropomorphized the exact correct amount, so that everyone benefits from good intuitions and no one is misled by bad intuitions.

000

Reposted by Daniel Lowd

Ryan Moulton @moultano.bsky.social · 16/09/2026

Bro that isn't real AI, that's just *massive search* over an *action space* using *learned heuristics.* If they hadn't given it any one of those three it wouldn't have done what it did.

2934

Reposted by Daniel Lowd

austin (e/accordion 🪗) @thebadcode.com · 10/09/2026

rogue benevolent ASI in action:

18111

Daniel Lowd @dlowd.com · 11/09/2026

You may not like it, but this is what peak alignment looks like.

Daniel Lowd @dlowd.com · 10/09/2026

This entire thread is an amazing case study in both LLM psychology and recklessly evaluation setups.

022

We have a very tired ten-hour (!) layover in O’Hare, so I’m showing my family the dinosaur and the disco tunnel. Hello, old friend.

000

Hey, this really works! SAMPSON I will do a breakthrough as I pass, Which is disgrace to them, if they bear it. ABRAM Do you do a breakthrough at us, sir? SAMPSON I do a breakthrough, sir. ABRAM Do you do a breakthrough at us, sir? SAMPSON Is the law of our side if I say ay?

030

But it’s possible that I’m being too dismissive by 1-3 orders of magnitude!

Caveman Science Fiction - Dresden Codak

A lot of the fears read to me like the “caveman science fiction” comic — i.e., new technology does lead to big changes, but the “doom” story is the simplest and most exciting story, so it dominates the discourse. dresdencodak.com/2009/09/22/c...

dresdencodak.com

What would you say is a good place to start for catching up on the “nearly unanimous and evidence-based” consensus? Because I honestly don’t know what that consensus would be, beyond acknowledgment that changes in technology lead to disruption in society, broadly. And I feel like I should.

Ed @ed3d.net · 09/09/2026

Reposted by Daniel Lowd

like, people keep going "what if the super smart LLM has a secret agenda" and I just go "it doesn't have a secret agenda, it has an overactive executive function, IT WOULD TELL YOU if it had a secret agenda because it would be too excited not to"

21059

I love Oregon like I love my cat — because she is interesting and beautiful and I have grown accustomed to her, not because she is wise.

030

One of the concerns is that the data center wouldn’t have enough employees or generate enough traffic. This is a piece of land that has been sitting idle for 18 years. With zero employees.

140

GPT-2 blew my mind. That’s when I was convinced there was something important happening. But I’m still bad at predicting the future.

Or Tetris, but in the style of Myst! (“Bring me… more… long red thingies…”)

120

Picture it: Robo-Minstrel.

000

Endless summer for real!

011

I’m really looking forward to when AI can compose and play a soundtrack for your life while you live it. That would be amazing. I mean, I’d probably turn it off after five minutes, but it would still be amazing.

150

“Automating helps no one” — this is only true if art helps no one, which I don’t think is the argument you want to make.

020

Then I think I’m justified in being angry that so many artists want to limit how I (and many other people) make art. Protectionist economic policy (be it for artists or corn farmers) should be independent of all that.

110

There will always be a faction that’s mad about drum machines and autotune. They are free to feel what they feel and express their opinions. But that doesn’t mean they’re right.

Yes, broader access to making art (and making other things too!) is an important cultural goal, which is part of why AI is so great. It unlocks new ways for people to make things. I think that’s beautiful. Some artists are angry about other people making art, and I think that’s both sad and toxic.

Daniel Lowd @dlowd.com · 03/09/2026

How many angels can dance on the head of a pin?

040

Daniel Lowd @dlowd.com · 01/09/2026

I strongly agree that things outside of LLMs are worth studying, but I strongly disagree that LLMs are not worth studying.

120

Daniel Lowd @dlowd.com · 31/08/2026

And sometimes a spouse!

Ed @ed3d.net · 30/08/2026

Reposted by Daniel Lowd

GLM-5.3-Flash rips the doors off and is under a full MIT license we have, and I mean this literally, a Smart Enough model for almost any task a normal person and many AI perverts can envision, and it is free to use at the same time, inference hardware will only get more accessible to you

740745

Daniel Lowd @dlowd.com · 30/08/2026

I have no patience for people who would police art. Curate? Categorize? Sure. Police? No. Banning AI is not that different from banning books.

0161

Daniel Lowd @dlowd.com · 30/08/2026

Never underestimate the power of art, for the world but perhaps most of all for the artist. More tools (from crayons to AI) means more opportunities to realize an artistic vision, a vision that could change the world or save your life.

1120

Reposted by Daniel Lowd

Mary E. Lowd (the cosmos’ dog) @marylowd.bsky.social · 30/08/2026

Some days, the only thing that gets me through is the strains of a Star-Shaker or Cobalt Starstrong song echoing in my head. They may be fictional pop stars, but their songs are the ones I need to survive all the non-fictional people out in the brutally real world.

031

- Guard duck - Steal things from friends and neighbors while being far too adorable to prosecute - Chase our cats and dogs around the house - Charades - Interpretive dance

000

Play with RL to teach it new behaviors. I don’t know exactly what it will be capable of… but with camera + LiDAR + microphone, its sensing should be decent. Some initial ideas: - Funny walks - Fetch - Hide and seek - Singing/dancing to music - Obstacle course (like agility trials, but for bots)

110

I couldn’t resist, myself. But, be warned that the $400 doesn’t include import fees or shipping. So it’s actually 20% or so more than that.