Tech Thread
- Thread starter MindzEye
- Start date
Cal Gal
Well-known member
"Thoughts for bedtime. . ."
palisaderesearch.org
Language Models Can Autonomously Hack and Self-Replicate
2026-05-07Language Models Can Autonomously Hack and Self-Replicate
We demonstrate that language models can autonomously replicate their weights and harness across a network by exploiting vulnerable hosts. The agent independently finds and exploits a web-application vulnerability, extracts credentials, and deploys an inference server with a copy of its harness...
palisaderesearch.org
MindzEye
Wayward Ditch Pig
axlsalinger
Well-known member
Significantly cheaper, though. Which may not lead to cheaper cars or better safety outcomes, but can I interest you in the world's richest man getting 0.001 % richer?
Cal Gal
Well-known member
D.C. Judge Keeps Heat On Google Over Search Data Sharing
Mehta Denies Pause On Google Data-Sharing Order
Judge Mehta denied a pause on Google’s data-sharing order; practical sharing is “months away.”
hoodline.com
Cal Gal
Well-known member
Canadian Watch-Party App Slaps Apple With Antitrust Suit Over App Store Ban
Canadian Watch-Party App Slaps Apple With Antitrust Suit Over App Store Ban
Rave sues Apple after App Store removal, seeks reinstatement and damages.
hoodline.com
How US automakers lobbied themselves into a $70B loss - and let China win
Automaker lobbying led to regulatory whiplash, making it harder to do their jobs and leading to $70B in losses - as the US falls behind China.
electrek.co
MindzEye
Wayward Ditch Pig
It's almost, almost like it's not about the economics and instead about social disruption or something
fortune.com
It's cool though, despite none of our previous predictions about how AI is going to scale in the future actually coming true, our new predictions about the cost of AI coming down "over the next 4 years" will totally come true. Trust me bro.
I'll say this now. If you like using LLM's for anything (I find them a pretty useful search engine/research aid for some things), look into self hosting one of the bigger Deepseek models. Apple is 100% right, locally run, personal LLM's are the future of this technology.
‘The cost of compute is far beyond the costs of the employees’: Nvidia exec says right now AI is more expensive than paying human workers | Fortune
Big Tech has announced $740 billion in capex this year, but AI has yet to show evidence of widespread increased productivity.
It's cool though, despite none of our previous predictions about how AI is going to scale in the future actually coming true, our new predictions about the cost of AI coming down "over the next 4 years" will totally come true. Trust me bro.
I'll say this now. If you like using LLM's for anything (I find them a pretty useful search engine/research aid for some things), look into self hosting one of the bigger Deepseek models. Apple is 100% right, locally run, personal LLM's are the future of this technology.