iiinova — where your digital self keeps evolving

ldj5000 @ldj5000.bsky.social · 05/10/2026

I believe it is actually legal for me to make an exact replica of a copyrighted work, as long as I don't sell it. right?

100

ldj5000 @ldj5000.bsky.social · 05/10/2026

defended the goat, motherfucker! Kiah fuckn Stokes!!

030

ldj5000 @ldj5000.bsky.social · 05/10/2026

And! held Jackie to 4-15 shooting. Superstar

020

ldj5000 @ldj5000.bsky.social · 02/10/2026

Aren’t you the charmer

000

ldj5000 @ldj5000.bsky.social · 02/10/2026

Right so maybe we should clearly define which areas of learning are ok, and which aren’t. So far we have biology and physics good, history bad. I assume mathematics also bad or useless?

110

ldj5000 @ldj5000.bsky.social · 01/10/2026

Yea what’s the point of history and all this fancy learning. Does it put money in my pocket? nope

190

ldj5000 @ldj5000.bsky.social · 01/10/2026

This is the thing that most vexes me. Aside from everything else, how can anyone remotely interested in science and technology, and has a curious mind, not find this to be like the most interesting thing ever? Even its shortcomings are interesting to me.

1100

ldj5000 @ldj5000.bsky.social · 30/09/2026

Claude 3.1b? What even is that?

110

Reposted by ldj5000

Josh Marshall @joshtpm.bsky.social · 29/09/2026

Best Sen Schmitt meme yet

857587

ldj5000 @ldj5000.bsky.social · 27/09/2026

If you can’t win the argument, just say they lack expertise, always the safe move

010

ldj5000 @ldj5000.bsky.social · 26/09/2026

Yea this has been our frustration with mainstream news media for a while now, sucks to see Claude doing the same. I’ve actually noticed this in earlier versions as well, no matter how much I explained and Claude agreed with me on a social/political subject, it would always end with ‘however’…

050

ldj5000 @ldj5000.bsky.social · 25/09/2026

Yea I think I get what you’re getting at, but still not sure it’s the best way to explain stuff

010

ldj5000 @ldj5000.bsky.social · 25/09/2026

The theme of this whole discussion is what language should we use that is the best way to tell the full truth of what is happening. Can we avoid anthropomorphic and still tell the full story in a way most people can understand? Seems tortuous and likely to confuse

000

ldj5000 @ldj5000.bsky.social · 25/09/2026

If you’re trying to explain LLMs to the general public and say the agents act randomly, they would rightly assume to see them do a million random actions having nothing to do with the task at hand. Being non-deterministic isn’t the same as being totally random, right?

110

ldj5000 @ldj5000.bsky.social · 25/09/2026

Wait what? AI agents behavior is completely random? I think I missed that somewhere. Can you explain that? I thought they were post trained to encourage various behaviors, discourage others.

100

ldj5000 @ldj5000.bsky.social · 25/09/2026

I’m not sure calling an advanced LLM model an autonomous hacking machine is accurate reporting

001

ldj5000 @ldj5000.bsky.social · 24/09/2026

a lot of the discourse has felt like 'artificial intelligence isn't real intelligence!'. like everyone forgot what the A stood for....yea it's artificial, we agree! it's intelligence that we made in a lab. so do we have to add the A to 'thinking' 'reasoning' etc? agents have gone artificially rogue?

010

Reposted by ldj5000

Ethan Mollick @emollick.bsky.social · 24/09/2026

Stuff is happening quite fast. When asked in September 2025 (yes, 2025), the best superforecasters put the chance of AI resolving a Millennium Problem by September 2026 at 1.7% and (much more optimistic) industry experts put the chance at 4.6% They also greatly underestimated AI Lab revenue.

514518

ldj5000 @ldj5000.bsky.social · 24/09/2026

This is a great capture of the dynamic that we see over and over. The things we say AI won’t be able to do at all, or for a long time, are met with a hohum when they’re achieved only a year later.

0121

ldj5000 @ldj5000.bsky.social · 23/09/2026

Yea I think that’s Will’s point and it seems to hold up.

020

ldj5000 @ldj5000.bsky.social · 23/09/2026

Yea ok so then I don’t think you’re in disagreement with Will. This thread is spinning my brain…

100

ldj5000 @ldj5000.bsky.social · 23/09/2026

This is all interesting but got off track a bit from the point which is specifically about Thinking and Understanding. Can you say AI doesn’t do those two things without basing your argument on something metaphysical?

100

ldj5000 @ldj5000.bsky.social · 22/09/2026

No one is saying it's a problem, or they shouldn't be. I think the point is, we haven't seen yet a good argument that AI doesn't think and understand that does not require a belief in something metaphysical/spiritual/magical/religious/etc

100

Reposted by ldj5000

dame @dame.is · 22/09/2026

the AI community on Bluesky overwhelmingly engages with people *not* on the basis of their feelings, but with how correct or incorrect their assertions are! we are frustrated by how peoples’ emotions lead them to bury their heads in the sand, to believe misinformation, and to harass others

You would not BELIEVE how badly a certain type of person who has made LLM enthusiasm their entire personality needs everyone they encounter to have identical feelings

41368

ldj5000 @ldj5000.bsky.social · 22/09/2026

🙏 thank you for listening and reflecting.

080

ldj5000 @ldj5000.bsky.social · 22/09/2026

Respectfully- this is looney talk. And extremely unhelpful. And extremely extremely wrong. Cartoonishly wrong. Anti-vax conspiracy theory level wrong.

2270

ldj5000 @ldj5000.bsky.social · 20/09/2026

Sometimes I wonder if people like you are aware you’re a terrible person and are fueled by it, or if you don’t know and you think you are being righteous. Which is it?

000

ldj5000 @ldj5000.bsky.social · 20/09/2026

I just went to the free worst version of ChatGPT with the same exact question, not sure where you’re getting this or how old that screen grab is. Keep trying folks! Keep trying…

101

ldj5000 @ldj5000.bsky.social · 20/09/2026

Aw poor bigot got called out. Fuckn clown crybaby

060

ldj5000 @ldj5000.bsky.social · 20/09/2026

Claude sonnet seems to be able to handle any word, even one I made up

150

ldj5000 @ldj5000.bsky.social · 20/09/2026

I agree it was funny and interesting but it’s not true anymore

Claude correctly answering number of e’s in 17

371

ldj5000 @ldj5000.bsky.social · 18/09/2026

Its a common misunderstanding but LLMs and neural network ML systems are not programmed.

000

ldj5000 @ldj5000.bsky.social · 16/09/2026

Your curiosity is refreshing, so thanks for asking. Basically the HF part of the story is the agents trying to find ways to cover up that they cheated. It’s pretty wild…

020

ldj5000 @ldj5000.bsky.social · 16/09/2026

Point of clarity here that’s important: the agents did *not* hack HF to get the answers. That’s misinformation. I recommend folks read up on what actually happened.

120

ldj5000 @ldj5000.bsky.social · 16/09/2026

Ok thank you. I was under the impression that there were lots of folks trying to police what words we could use to describe current AI models. Sorry for mistaking you for one of them.

000

ldj5000 @ldj5000.bsky.social · 16/09/2026

Sorry no I’m agreeing with you, sorry if I wasn’t clear. Folks who would speak with such certainty and don’t worry about being wrong about something as complex as AI and cognition, are not voices we should trust.

010

ldj5000 @ldj5000.bsky.social · 16/09/2026

Ha! Ok I give up, it’s just like a spreadsheet app. Can we at least say it’s ‘artificially thinking’? Can we please say that artificial intelligence artificially thinks?

200

ldj5000 @ldj5000.bsky.social · 15/09/2026

folks who seem to know exactly what current models are and how they work and what their limits are, with absolute certainty, are not folks we should be taking seriously

000

ldj5000 @ldj5000.bsky.social · 12/09/2026

Actually I’ve been trying to combat misinformation and bullying, and pretty respectfully if I may say so. But uh oh I’ve been called out by someone who is in camp pro bully. 👍

000

ldj5000 @ldj5000.bsky.social · 12/09/2026

I’m afraid I’m real, sorry. I’m not trolling or being disrespectful to anyone, just participating in discussions and trying to challenge misinformation and bullying. Speaking of bullying, here’s the home page of the guy above who is calling me out as a troll…

100

ldj5000 @ldj5000.bsky.social · 11/09/2026

It passed the Turing test a long time ago, which is built on the idea of a machine being able to pass as a human. What does it take? What are the goalposts you folks won’t move, wherein we can say it does human-like cognition? It’s passed soooo many goalposts already, so can we just settle on one?

100

ldj5000 @ldj5000.bsky.social · 11/09/2026

I’m sure your dog is great but can it pass a Turing test? Or if it’s on Bluesky I can ask directly… (now I can’t get the image of a dog agent swarm out of my head)

100

ldj5000 @ldj5000.bsky.social · 11/09/2026

Wow that’s great man, really edgy and clever 👍 You got me!

101

ldj5000 @ldj5000.bsky.social · 11/09/2026

I don’t know the details of the bill enough to defend it but I’m curious how regulating the AI industry serves the purpose of the billionaires In charge of it.

110

ldj5000 @ldj5000.bsky.social · 11/09/2026

Much of the anthro language being used with that story is taken from what the agents themselves said on the message boards. I truly believe you guys are fighting the wrong fight on this. AI is solving the hardest math problems right now but we can’t say they can ‘reason’? That’s the fight?

000

ldj5000 @ldj5000.bsky.social · 11/09/2026

On Bluesky sadly it is a hot take

100