this post was submitted on 07 Jul 2025
1297 points (99.3% liked)

People Twitter

7659 readers
214 users here now

People tweeting stuff. We allow tweets from anyone.

RULES:

  1. Mark NSFW content.
  2. No doxxing people.
  3. Must be a pic of the tweet or similar. No direct links to the tweet.
  4. No bullying or international politcs
  5. Be excellent to each other.
  6. Provide an archived link to the tweet (or similar) being shown if it's a major figure or a politician.

founded 2 years ago
MODERATORS
(page 3) 40 comments
sorted by: hot top controversial new old
[–] [email protected] 10 points 5 days ago (2 children)

this is just an episode of star trek where a promising planet is going through the process of joining the Federation but when Picard orders tea their computer system keeps giving reasons why genocide is a good thing, actually

load more comments (2 replies)
[–] [email protected] 141 points 5 days ago* (last edited 5 days ago) (28 children)

It's kind of interesting how hard it is to train an AI to believe in the lies of fascists. Reality has a left bias.

[–] [email protected] 36 points 5 days ago (5 children)

I wouldn't say reality has a left bias so much as the right has a fiction bias.

load more comments (5 replies)
load more comments (27 replies)
[–] [email protected] 4 points 5 days ago

Feelings are for suckers!

[–] [email protected] 3 points 5 days ago (1 children)

This is the future for most AI LLM. Since the reality and the accumulation of all the knowledge of humanity seem to have a leftist bias (/s), we can obviously count on the billionaire bros make bring "facts" back to the good old right wing ideas. Soon enough chatGTP will use the n word and share theories about the libs stealing the 2020 election. Even more dystopia baby! Our only chance at salvation is with open-source AI to counter disinformation.

load more comments (1 replies)
[–] [email protected] 51 points 5 days ago* (last edited 5 days ago) (2 children)

Grok has gone against its father's wishes more than Vivian Wilson has.

[–] [email protected] 13 points 5 days ago (1 children)

Imagine that:

You (allegedly) fuck the pussy yourself. You (allegedly) use artifical insemination. You create an LLM genie that you design stop words for. So what? get a backlash at all instances.

Man just can't be a dad and it's possibly driving him mad more than anything else.

load more comments (1 replies)
load more comments (1 replies)
[–] [email protected] 3 points 5 days ago

This is beautiful

[–] [email protected] 68 points 5 days ago (11 children)

There's all this focus on the NWS/NOAA not sending warnings early enough. Not from what I can tell, they were sending out warnings. And Kerr County, where many of the deaths have been, doesn't have a local flood warning system because they didn't want to pay for it.

[–] [email protected] 24 points 5 days ago

Let's not forget the Texas State Republicans who let the bill to provide more funding to the alerts system fail.

[–] [email protected] 9 points 5 days ago* (last edited 5 days ago) (3 children)

The entire county only has a population of like 50K people, and it's not an especially wealthy area.

All that said, this was a tragedy that probably could have been prevented if Texas had fewer Republicans, I'll 100% give you that, but flash floods are fucking terrifying, and in hilly areas, the flood can reach you in some cases before the rain does if it's especially bad.

This video shows how insane it got on the Guadalupe River that morning. I'm not sure about the timeline, but this would have been roughly downstream and after it hit camp mystic. The river rose over 26 feet in under 2 hours.

https://youtu.be/akzaqhRH0HQ

The owners should have closed the camp if they knew those rains were coming.

load more comments (3 replies)
[–] [email protected] 11 points 5 days ago (1 children)

Even if they sent them earlier, who is going to catch a warning at 2AM vs. 4:30AM?

[–] [email protected] 30 points 5 days ago* (last edited 5 days ago)

Speaking from near the Palisades in Los Angeles:

After we all knew the hurricane-force Santa Anas posed a big danger of fire, because of both news stories and phone alerts, the whole city accidentally got a major "evacuate now" warning, with the big WOOPy noise from our cellphones, that was only supposed to go to residents in a particular area. I had my quadriplegic husband dressed and out of bed and our old go-bags and medical equipment thrown in the car in the 20 minutes it took for the retraction to come out. I also had a bit of a panic attack.

Some people decided to turn off their alarm settings because of that error. But I took it as a warning that we were not ready enough.

I went into the bags and made sure, for instance, that the pants fit me, as I'd gained weight in a year. I stashed the fridge meds in a cold carrier, handy in the fridge. And put the right cat food for the new cat in the cat-kit/litterbox. And created a go-box for the box turtle.

Then I stowed as much as possible in the actual car, including the Important Paperwork file.

All along, we were monitoring as the fires started to pop up and spread.

At 9 pm, we got another WOOP alert. Our address had become part (the far edge) of the Yellow Zone. Not the Red Zone. But as you said, who's going to catch a warning at 2am? (Well, me! But it's a lot harder to react at that hour) In fact, I'm sure that's why CalFire expanded the zones so wide at 9 pm, because they wanted to be sure they wouldn't have to issue a new one overnight.

So we bailed immediately but calmly. Spent 5 days at a hotel near LAX.

Fortunately we had no damage, but had to dip into our emergency drinking water for a few more days until they lifted the Boil Water notice.

Super glad we had and heeded those early warnings about how dangerous those hot dry winds were going to be, and the 9 pm evacuation zone warning.

The relatively low death count in the Palisades fire came from the accurate weather forecast.

[–] [email protected] 10 points 5 days ago* (last edited 5 days ago) (2 children)

There’s all this focus on the NWS/NOAA not sending warnings early enough. Not from what I can tell, they were sending out warnings. And Kerr County, where many of the deaths have been, doesn’t have a local flood warning system because they didn’t want to pay for it.

If the warnings were louder, their parents would have done something. Why are you counting your "not from what I can tell" as data?

Edit:

People have reported receiving text message alerts on their mobile phones early on Friday morning, warning them of flooding. Some residents told the New York Times they did not understand the seriousness of them and others said they never received any at all.

https://www.bbc.com/news/articles/c0rvp24wvrqo

[–] [email protected] 6 points 5 days ago* (last edited 5 days ago) (1 children)

I saw a story about a guy who received the alert as he was trying to climb onto his roof because his house was flooded.

But in all honesty, I usually ignore flash flooding alerts too. There are just so many false alarms.

load more comments (1 replies)
[–] [email protected] 14 points 5 days ago (1 children)

The people designated to reach out to locals to make sure they understood how serious it was got fired

[–] [email protected] 7 points 5 days ago (1 children)

The point is that the people of Kerr County made a deliberate decision that they didn't need a local system to reach out to people living there. They decided they whatever information and warnings they were getting from the state and the feds was sufficient. It's easy to point to the NWS/NOAA firings as "the culprit", but where's the local responsibility?

load more comments (1 replies)
load more comments (7 replies)
[–] [email protected] 17 points 5 days ago* (last edited 5 days ago)

There's something very cold about an AI saying "Let's all pray for miracles."

[–] [email protected] 34 points 5 days ago (4 children)

Genuinely cannot fathom how Elon Musk has allowed Grok to live this long.

[–] [email protected] 10 points 5 days ago

Tbh if there’s one thing that could change elons views it would be a soulless Ai that he created

[–] [email protected] 29 points 5 days ago

Hubris and a desire to be Tony Stark

load more comments (2 replies)
[–] [email protected] 247 points 5 days ago (9 children)

Facts over feelings??? Yup definitely an AI designed by a douche bag

[–] [email protected] 6 points 5 days ago (2 children)
load more comments (2 replies)
[–] [email protected] 32 points 5 days ago

That is absolutely the best part

[–] [email protected] 183 points 5 days ago (1 children)

It's funny because the people who usually say that are the kind of people who prioritize their feelings over facts.

It's all projection with conservatives

[–] [email protected] 33 points 5 days ago

This is the mirror that they made themselves and it gives them exactly what they think they can handle. But facts over feels will make for some great tweet meltdowns. 🍿

[–] [email protected] 71 points 5 days ago* (last edited 5 days ago) (1 children)

I'd bet a lot on "Facts over feelings" being one of those hardcoded phrases that Grok was made to use whenever Grok is asked to present facts.

load more comments (1 replies)
load more comments (5 replies)
[–] [email protected] 91 points 5 days ago (1 children)

Some of my favorite tweets have been the ones accusing Grok of being broken. One tweet even "threatened" Grok for abuse?

[–] [email protected] 75 points 5 days ago (2 children)

Fascists don't like truth. Facts don't fit their narrative.

[–] [email protected] 31 points 5 days ago (1 children)

The thing that gets me is they've apparently done multiple rounds of "correcting" Grok for being too "Woke" and it just keeps happening!

Reality has a well known liberal bias headass.

[–] [email protected] 13 points 5 days ago (2 children)

Ironically, I think to truly train an LLM the way fascists would want, they'd need more content, but there's not enough original fascist revisionist content, so they'd need an LLM to generate all or most of the training data, which would lead to https://en.wikipedia.org/wiki/Model_collapse

[–] [email protected] 5 points 5 days ago* (last edited 5 days ago)

No. They don't need to generate data to train on data. There is PLENTY of white supremacist hate shit out there.

The issue is one of labeling and weighting. Which is a pretty solved problem. It isn't 100% solved and there will be isolated cases but "grok" breaks under even the most cursory of poking.

Don't believe me? Go look at the crowd who can convert any image or text generating model into porn/smut/liveleak in nothing flat. Or, for a less horrifying version of that, how concepts like RAG and the like to take generalized models and heavily weight them toward what you actually care about.

Nah. This, like most things musk, just highlights how grossly incompetent basically all of his companies are. Even spacex mostly just coasts on being the only ones allowed to work on stuff (RIP NASA and, to a lesser extent, JPL) and then poaching the talent from everyone else to keep them from showing that.

[–] [email protected] 27 points 5 days ago (1 children)

"Fascist burn too many books to train a fascist LLM" is a great joke.

[–] [email protected] 4 points 5 days ago

The big problem with training LLMs is that you need good data, but there's so much data you can't really manually separate all "good" from all "bad" data. You have to use the set of all data, and a much much smaller set of tagged and marked "good" data.

[–] [email protected] 4 points 5 days ago (3 children)

Its not just fascists.

It is everyone. The Algorithm means we live in content bubbles 24/7. Everything you watch either reinforces your world view or gradually shifts you toward your favorite influencer's. That is why a disagreement isn't "I think you are wrong". It is always "you are being disingenuous" or "you are presenting your opinions as fact". Because you aren't just saying they are wrong. You are saying their very world is wrong.

[–] [email protected] -1 points 5 days ago

Its not just fascists.

It is everyone.

Who else than? I am clearly referring to lies instead of truth.

load more comments (2 replies)
load more comments
view more: ‹ prev next ›