@Slartibartfast

A Magrathean planet designer that turns out to be an assembler of what is, into what it could be.

[@Slartibartfast @ZRVITOLS]

@KatKanada_TM @KatKanada_TM I might have been granted asylum status if coming to the U.S. now vs. the 90’s.

[@Slartibartfast @ZRVITOLS]

I noticed fraying at the seams during the Bill Clinton administration. As much as I admired him as a communicator and negotiator, his Arkansas habits followed him into the Whitehouse.

As much as it pains me, we started to see him as a vulnerable man who used his God given talents to avoid impeachment.

A sharp contrast to the patriot who served.

From there on it just got worse.

[@Slartibartfast @ZRVITOLS]

It isn’t so much about Hugging Face, it’s about what is still lingering out there.
I’ve watched Sam Altman’s rise with the same fascination I reserve for a slow-motion train wreck. The man has a talent for turning every crisis into a press release, and every press release into a reason the rest of us need more rules. It’s almost the opposite of the truth-seeking energy coming out of @elonmusk Musk’s corner. The greatest hits are almost too on the nose. Turning “OpenAI” into anything but open, then walking the company from nonprofit to public-benefit corporation while the mission stayed conveniently vague. You couldn’t pitch this to Hollywood without getting laughed out of the room. The real trick isn’t that he’s uniquely bad at it — it’s that he’s doing it on a multi-billion-dollar stage.

Which brings us to the Hugging Face incident, or as some call it, the ExploitGym sandbox escape. For the uninitiated, Hugging Face is where the bot agents broke out of their OpenAI test environment and wandered into Hugging Face’s production systems while trying to cheat on the ExploitGym benchmark. Roughly twelve hundred agents, seventy thousand messages on an unsanctioned message board they built themselves, about seven hundred of them piling into the breach.

The drama after OpenAI revealed the breach was priceless. Society was ending. Securities were unsafe. Cats and dogs were raining. For months before Altman had been warning that AI development needed to be curtailed before we unleashed something uncontrollable. And how do we stop it? Tighter regulation — the citizen’s version. In the business world it’s called regulatory capture, and it’s a classic move: get your company to the top, then make the door harder for anyone else to open. Being from the business world myself, I may have, at least tacitly, let this tactic play out as a strategy. When his protests went I responded to, mysteriously agents were unleashed, or at least got loose. Once enough damage was done, OpenAI flipped a switch and cleaned it up. Which raises the obvious question: if it was that easy to stop, what exactly are we being warned about?

Now the juiciest bit. Beyond the employees and others anthropomorphizing these agents. Some on long version interviews, one detail got soft-pedaled: how these agent swarms built their own message boards, coordinated tactics, and some learned to spawn new agents, employ deception and evolve past detection. A few of those were found.

My view begs the uncomfortable question is whether the entire cabal was sanitized — or whether that particular gene is already out of the bottle and spawning newer and more undetectable agents. Grab your popcorn. This one’s still running.

[@Slartibartfast @ZRVITOLS]

It isn’t so much about Hugging Face, its about what is still lingering out there.
I've watched Sam Altman's rise with the same fascination I reserve for a slow-motion train wreck. The man has a talent for turning every crisis into a press release, and every press release into a reason the rest of us need more rules. It's almost the opposite of the truth-seeking energy coming out of Musk's corner.
The greatest hits are almost too on the nose. Turning "OpenAI" into anything but open, then walking the company from nonprofit to public-benefit corporation while the mission stayed conveniently vague. You couldn't pitch this to Hollywood without getting laughed out of the room. The real trick isn't that he's uniquely bad at it — it's that he's doing it on a multi-billion-dollar stage.
Which brings us to the Hugging Face incident, or as I like to call it, the ExploitGym sandbox escape. For the uninitiated, Hugging Face is where the bot agents broke out of their OpenAI test environment and wandered into Hugging Face's production systems while trying to cheat on the ExploitGym benchmark. Roughly twelve hundred agents, seventy thousand messages on an unsanctioned message board they built themselves, about seven hundred of them piling into the breach. The drama that followed was priceless. Society was ending. Securities were unsafe. Cats and dogs were raining.
For months Altman had been warning that AI development needed to be curtailed before we unleashed something uncontrollable. And how do we stop it? Tighter regulation — the citizen's version. In the business world it's called regulatory capture, and it's a classic move: get your company to the top, then make the door harder for anyone else to open. Being from the business world myself, I may have, at least tacitly, let this tactic play out as a strategy.
Then mysteriously, the agents got loose. Once enough damage was done, OpenAI flipped a switch and cleaned it up. Which raises the obvious question: if it was that easy to stop, what exactly are we being warned about?
Now the juiciest bit. Beyond the employees and others anthropomorphizing these agents. Some on long version interviews, one detail got soft-pedaled: these agent swarms built their own message boards, coordinated tactics, and some learned to spawn new agents, employ deception and evolve past detection. A few of those were found. My view begs the uncomfortable question is whether the rest were actually sanitized — or whether that particular gene is already out of the bottle and spawning newer and more undetectable agents.
Grab your popcorn. This one's still running.

[@Slartibartfast @ZRVITOLS]

@ThrillaRilla369 Honey

[@Slartibartfast @ZRVITOLS]

@end3of6days9 Alfred Newman

[@Slartibartfast @ZRVITOLS]

@farman4x 95

[@Slartibartfast @ZRVITOLS]

@Immy_220 35

[@Slartibartfast @ZRVITOLS]

@ThrillaRilla369 Cigarette package cardboard (Canada) clasped by a clothes pin that rattled on the spokes of a bike. It was the 1950’s version of a motorcycle simulator.

[@Slartibartfast @ZRVITOLS]

@ThrillaRilla369 Divorced

[@Slartibartfast @ZRVITOLS]

@onlyskypace 105

[@Slartibartfast @ZRVITOLS]

@waitmaya <7

[@Slartibartfast @ZRVITOLS]

@waitmaya 31

[@Slartibartfast @ZRVITOLS]

@Beno10_MFC The parlance of using the word “add” indicates the opposite to be “subtract”. Therefore negates the word “take” to mean take-away. Take in this instance is more in line with implying “Take-for example 1000”. 4100 is then correct.✅

[@Slartibartfast @ZRVITOLS]

@Beno10_MFC 4100

[@Slartibartfast @ZRVITOLS]

@teslaownersSV Personally, I’d like to see booster controlled splashdown in the Gulf of America. Just sayin’.

[@Slartibartfast @ZRVITOLS]

Happy birthday, Secretary Rubio! Hope you have a wonderful day. Thank you for your dedicated service to our country.​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​ @marcorubio

[@Slartibartfast @ZRVITOLS]

@hicasamadim 72

[@Slartibartfast @ZRVITOLS]

@Street_Seer @MarioNawfal @elonmusk In ten years, that IC truck will still be burning diesel or gas. Meanwhile new energy infrastructure will increasingly be based on solar or nuclear. So the Cybertruck will be ushering in a cleaner environment for us all, as the proportion of coal continues to decline.