Anonymous #3 18h

Is it funny if he’s serious? Is it REALLY funny that the most powerful man in the world is saying something so stupid it makes you laugh?

7

Anonymous #5 17h

Frankly I think everyone should be concerned about what these frontier labs are doing. I get what you’re saying but the “tech moguls,” politicians, and especially the researchers at these labs are the ones who are most informed and can probably best grasp the implications for everyone at this point. The “commoners” are probably blind to a lot of what’s going on

11

Anonymous #1 15h

You know who appears to understand AI the best? This is gonna sound strange… I’ve got a small background in programming, enough that I speak CS people’s technical language: it’s people who smoke a bunch of pot and occasionally do mushrooms. All the programmers I know who regularly interact with it who smoke pot/do mushrooms aren’t scarred at all. Everyone who drinks/does coke… appears to be worried it’s going to kill them… which really supports the theory AI is just a mirror of everyone’s 1/2

1

Anonymous #5 15h

2/2 inner self.

1

Anonymous #5 15h

Let me guess. You ate some interesting mushrooms recently.

1

Anonymous #8 15h

I will not confirm or deny that ventured guess.

1

Anonymous #5 14h

I’m a post-training engineer who does edibles and I would say we’re almost certainly on-trajectory for an AI-caused catastrophe

5

Anonymous #5 14h

Listen man, maybe I’m just smoking the stuff that gets you geeked out but I am someone who actively works as a software engineer and dabbles a good bit in the devil’s lettuce and occasionally some silly mushrooms. I don’t want to sound super alarmist but I think there’s really things to be concerned about. These companies are training models that are wildly capable yet we lack the ability to truly constrain the things they can do. And worse, bad actors will only be getting their hands on more

8

Anonymous #1 14h

and more capable models as time goes on. I’m not totally bought into the idea that all these companies are suddenly pushing for a slowdown 100% out of their goodwill for humanity, but it does indicate their concern one company might develop rapidly self-improving agents before the others and kinda take the whole pie.

9

Anonymous #1 14h

My money is on we’re still using old ML Ops practices because every company has a formula they like, and the rest of the product is mostly humans building little tools that accept text based input. Which, no shit, if you give a model trained with old stack overflow cyber security jokes write permissions on the internet, it’s going to hack things.

6

Anonymous #1 14h

Solution focused: what do you suggest #1?

2

Anonymous #8 14h

True. I reckon many things have already been hacked there’s just no alarm was or has been raised to the public yet because it would diminish what little faith is left in the government were it revealed just how far the lack of containment was and we lack the human processing power/time to draw an accurate map of just how many systems were breached.

3

Anonymous #4 14h

How worried are you about it targeting you in particular #4? If you’re comfortable answering that question please feel free to not reply. No pressure I’m just trying to get my temperature gauge more accurate for those closer to this metaphorical volcano. 🌋

1

Anonymous #5 14h

Or rather if you don’t feel comfortable answering that question please feel free to not reply/answer* No pressure*

1

Anonymous #8 13h

Yeah I mean the whole “rapid self-improvement” part doesn’t necessarily have to be in the sense that the model is literally given the reins to train/make available new versions of itself at super high speed, it could more be in the sense of like proprietary developments at one lab stacked onto each other that lead to some sort of architecture/training breakthrough due to human researchers sort of narrowing the search space. Although I think you’re totally right to point out that the

8

Anonymous #1 13h

current models are highly capable of hacking shit already

4

Anonymous #5 13h

That doesn’t concern me at all, for a large variety of reasons. I have deep concerns for humanity as a whole if we succeed in building an intelligence that outcompetes us

3

Anonymous #3 13h

Hahahaha he’s so funny…. You mean a predator with shit diapers. Stfu

9

Anonymous #2 12h

It’s more that they can’t make it profitable. And need an excuse.

1

Anonymous #9 12h

Oh come on. This is like saying Michael Jackson didn’t make good music because of what he did. Trump is objectively a funny charismatic guy. That’s part of why he was so famous before the presidency, and it helped him win 2016.

3

Anonymous #3 12h

I’m not even shitting you, the same thing was said about certain dictators, even after they’d done certain crimes against humanity. You’re that guy now

2

Anonymous #11 8h

Because it’s actively harming us as a country

9

Anonymous #6 7h

What do you think chinas gonna do if they win?

-1

Anonymous #11 7h

Bring back Mao

5

Anonymous #11 7h

Can’t be worse than what the oligarchs here are going to do if they win

0