this post was submitted on 23 Apr 2026
177 points (91.9% liked)

Technology

84199 readers
4276 users here now

This is a most excellent place for technology news and articles.


Our Rules


  1. Follow the lemmy.world rules.
  2. Only tech related news or articles.
  3. Be excellent to each other!
  4. Mod approved content bots can post up to 10 articles per day.
  5. Threads asking for personal tech support may be deleted.
  6. Politics threads may be removed.
  7. No memes allowed as posts, OK to post as comments.
  8. Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
  9. Check for duplicates before posting, duplicates may be removed
  10. Accounts 7 days and younger will have their posts automatically removed.

Approved Bots


founded 2 years ago
MODERATORS
all 43 comments
sorted by: hot top controversial new old
[–] Hegar@fedia.io 111 points 5 days ago (2 children)

Lol the gizmodo article about the preview of mythos has this incredible openly line:

If Anthropic is kind of like a kid at a Halloween party with a bowl full of spaghetti, telling its blindfolded friends the spaghetti is brains, European cyber agency leaders are like kids who are mad that they still haven’t gotten to stick their fingers in the brain bowl

Anthropic in particular seems to use "we're scared of what we can do" as their marketing strategy. So maybe this unauthorized use is real, or maybe it's just more PR.

[–] cecilkorik@piefed.ca 74 points 5 days ago (1 children)

Narrator: "It was more PR."

[–] elvith@feddit.org 15 points 5 days ago (1 children)

Chill bro! It was all a pr... stunt!

[–] YiddishMcSquidish@lemmy.today 7 points 5 days ago* (last edited 5 days ago)

::points at ~~camera~~ piles of cash:: chill bro chill bro! It's for the ~~gram~~ money.

[–] SaveTheTuaHawk@lemmy.ca 1 points 4 days ago

They are generating a bullshit PR story every week about how dangerous and awesome Mythos is.

[–] Willdrick@lemmy.world 35 points 5 days ago (2 children)

Seems more and more transparent that the only strategy Dario is pushing is to scare people so governments ban FOSS AI, that way most people would have to use "the safe sanctioned AI" (and pay through the nose for it).

I wonder if politicians getting lobbied are greedy as always or just retarded.

[–] very_well_lost@lemmy.world 11 points 5 days ago

I wonder if politicians getting lobbied are greedy as always or just retarded.

Yes.

[–] Willdrick@lemmy.world 2 points 5 days ago

Seems I should be careful with my language. I seem to have offended a politician

[–] treadful@lemmy.zip 51 points 5 days ago (1 children)

But don’t worry, this secret group that apparently has access to the most feared piece of technology in the world is “interested in playing around with new models, not wreaking havoc with them,” the source apparently explained to Bloomberg.

Not sure I believe this shtick anymore. There's always going be at least that one person in the group that can't handle their shit.

They could also just revoke their access no? I mean its not like they actually release the model files do they?

[–] RIotingPacifist@lemmy.world 44 points 5 days ago* (last edited 5 days ago) (2 children)

~~OpenAI~~ Anthropic: Be scared we have developed AGI

Also ~~OpenAI~~ Anthropic: our permissions model is dogshit, we just wildcard match strings

[–] garbage_world@lemmy.world 15 points 5 days ago (1 children)

Mythos is made by Anthropic, not OpenAI.

[–] RIotingPacifist@lemmy.world 17 points 5 days ago (1 children)

Is there a meaningful difference?

[–] rangber@lemmy.zip 15 points 5 days ago (1 children)

One has Mythos, the other doesn't for now

Anthropic is doing what OpenAI did for chat GPT5.

[–] Iconoclast@feddit.uk 0 points 5 days ago (1 children)

They haven't claimed to be in the possession of AGI no matter how nicely that would fit your narrative.

[–] XLE@piefed.social 10 points 5 days ago* (last edited 5 days ago)

You are simply wrong. Anthropic specifically has been caught extensively lying about their chatbot capabilities.

An Anthropic executive claimed his chatbot had feelings (as an excuse to harass LGBT people).

[–] BigTrout75@lemmy.world 27 points 5 days ago (3 children)

Would be funny if it was the Pentagon

[–] prex@aussie.zone 38 points 5 days ago (5 children)

Funny haha or funny torment nexus?

[–] popekingjoe@lemmy.world 16 points 5 days ago

A little of Column A, a little of Column B.

[–] DevDave@piefed.social 6 points 5 days ago

They can't make it to obvious this is the bad place.

[–] Hegar@fedia.io 6 points 5 days ago

If there's one thing fascism has taught us, the two are hardly mutually exclusive.

[–] trackball_fetish@lemmy.wtf 1 points 4 days ago

Torment nexus really is an info hazard, not because the ai will make it so but because humans are fucking evil and dumb enough to love the idea

[–] cecilkorik@piefed.ca 2 points 5 days ago

I wish we had judged "Don't create the torment nexus" by its cover.

[–] Cruxifux@feddit.nl 4 points 5 days ago* (last edited 5 days ago)

I never even thought of that. After hearing about this I always just assumed some private entity or hacker group or billionaire scumfuck was who this “secret entity” was but it absolutely could be, and is likely to be some world power government entity.

[–] DeckPacker@piefed.social 11 points 5 days ago

It's not real, please don't fall for this obvious marketing bullshit.

[–] stopdropandprole@lemmy.world 11 points 5 days ago (2 children)

why are they hosting it on the fucking Internet? I just don't understand the lack of concern for safety. a few years ago leading experts said AI is an existential threat, but I guess after enough rounds of investor funding, the risks magically all disappeared?!

[–] SaveTheTuaHawk@lemmy.ca 17 points 5 days ago (1 children)

It's all bullshit. Some Anthropic engineer told a story about how a sandboxed computer was told to try and connect to the internet and it sent him an email during lunch.

Stories to convince investors they have technology that can one day tell you how many R's are in strawberry, or why the alphabet has two Ps.

[–] _haha_oh_wow_@sh.itjust.works 5 points 5 days ago (2 children)

"Soon we will be able to reliably set a timer!"

-Some asshole saying this unironically

[–] Doomsider@lemmy.world 4 points 5 days ago

I am pretty sure he said that the timer was still a year out.

Meanwhile, I picked up a stop watch at a second hand store for a dollar.

[–] QuandaleDingle@lemmy.world 1 points 4 days ago (1 children)

I don't know why a LLM needs this feature. Just...use an assistant?

[–] SaveTheTuaHawk@lemmy.ca 2 points 4 days ago

You missed the point.

[–] sturmblast@lemmy.world 1 points 4 days ago

Hype, stock price go up

[–] inari@piefed.zip 11 points 5 days ago (3 children)

I love when people use this photo of Dario for bad news about Anthropic

😬

[–] Iconoclast@feddit.uk 3 points 5 days ago (1 children)

The photos of people that news outlets use reliably reveal their bias. If they don't like the person they use a bad picture. Once you notice it, it can't be unseen.

[–] XLE@piefed.social 4 points 5 days ago

Iconoclast, what's wrong with casting creeps as creeps? The guy in the picture enthusiastically helped the Department of Defense kill Iranian schoolchildren.

IIRC, weren't you also defending the AI creep/abuser Eli Yudkowsky too?

[–] dan69@lemmy.world 5 points 5 days ago

I knew it. There was never a locked down on only those “few” ppl who have access!!

[–] melfie@lemmy.zip 6 points 5 days ago

Cool, let me know when the model leaks. 🥱

Bro I signed up for openai's ai when it wasn't a thing yet 2years prior to chatgpt and they where saying its dangerous so only researchers get access to it same as mythos. I entered literal gibberish bullshit into a form and ticked a box promising not to use it for bad things and they gave me access so I doubt anthropic are going to be any more secure.