How Did This Happen??

Its 2030, I turn on my leased iPhone to open up a tech linked video, get a notification for an overdraft due to my YouTube Premium charge and get locked out of my phone due to the Ai detecting CSAM (it was a photo of my nephews birthday). Due to legal obligations on tech companies reporting to governments over detected violations I am now sitting on my front step waiting for the police to arrest me because my appeal to the ai was reviewed by another ai and rejected. The officer pulls up without his body cam as he now has meta raybans. The rayban video ends up in a Meta data leak and now my video is on YouTube, so I go to defend myself in the comment sections because after a judge review and thousands of dollars in lawyer fees I was able to prove my innocence. But my premium subscription didn't go through so now I have to sit through 3 unskippable ads to read the publics opinion (which are mostly astro turfing bots). I then try to comment and immediately get my comment removed due to youtubes auto moderation deleting my comment but not the profile with a naked woman in its pfp. My comment gets flagged for csam because the LLM was trained on the leaked data from meta, so now I am flagged and reported to the authorities again. Cant wait for the future.

I believe nothing Altman says. He's trying to replicate the PR that Mythos had for his model, because OpenAI is in freefall.

"With no Internet access" ah so the AI grew legs

I hope to live long enough to see Sam Altman living in a cardboard box in San Fransisco's Tenderloin District...

Man, I better start sailing the high seas now that I can use the AI defense. "I pirated movies, but I was using it to train myself to better my filmaking skills, therefore it's fair use."

So now public listed companies officially hacking data and publically admitting data theft... Brilliant ...just brilliant 😂

Why does this hugging face hack sound like a fake press release to promote their newest model

Fining Anthropic $3k/book for piracy is like fining a thief the cost of dinner for clearing out a jewellery store.

It escaped a "closed sandbox with no internet"? Or some researcher broke in and they blame it on the bot to avoid jail time?

All I heard from the anthropic case is that I could buy a game or software of any kind, decompile it, and use it however I like.

I bet you 100% that the open AI model did actually have internet access from the get go and they're lying OR they ran it in a virtual machine where the host machine had internet access but the VM didn't...which is a goddamn oversight since that should have been taken into consideration.

Apple partnering with Klara just shows you how expensive the iphone 18 is going to be.

"Eucharist of the second coming will not be born of bread and blood, but on a substrate of silicon dioxide. Despair, denizens of Technus, for all written knowledge will be transubstantiated into the churning, insatiable maw of our new computer lord. A god who has no idea how many Rs there are in strawberry" - Grand Archduke Riley Murdock of Technus

Techlinked is actually a multi-year-long project documenting Riley's slow downward spiral of despair as the tech industry become more and more dystopian.

0:19 what is the chance that this is just a publicity stunt by openAI?

remember sam altman wants open source models banned

1:32 is anyone else starting to realize that this whole “Ai breaching containment” thing just sounds like an excuse to keep AI closed source? It sounds like a completely fabricated lie perpetrated by tech bros in order to keep the unregulated tech oligarchy going. When Chinese AI releases with matching or exceeding parameters months or weeks later, they never report breaches at the same frequency as western Ai companies. This proves that Ai can be optimized with more regulation, less energy use and less breaches; this is a better approach to Ai (in my opinion) rather than the brute force approach companies like open AI have been using, leading to more breaches and wayyyy more energy use. But what do I know, I’m just an outsider looking in. All we can do now is just wait for new open source AI to see if it’ll really breach containment or if it’s just hot air.

These were a couple of PR advertisements based on deliberately created incident, meant to get across two things primarily: that the open models must be allowed to exist without guardrails because otherwise there would have been no defense possible (hugging face's point that the closed models with guardrails didn't even work), and that the closed models must be allowed to flourish with minimal guardrails because that's when they demonstrate their true power (Open AI's point, their model only escaped when they loosened the guardrails a bit). So forget alignment, it's full speed ahead on implementation if the United States is going to stay ahead on all of this. The two groups admitted they're working together to resolve this issue, my point is perhaps they worked on it together to make it happen. There's no denying this incident served both their interests, and all the AI companies have been working from this same playbook from the beginning. So the danger isn't really about what happened, it's about what happened to let it happen.

Escapes a closed virtual system with no Internet. Yet the hardware was connected to the Internet. Not a closed system.

i am going crazy, i could swear he says ban kids under 50 at 8:16 lol