@mhdksafa Because billionaires give millions to politicians and working-class immigrants don't.
Politicians follow the money, and the sheep just follow their masters.
ok it's time for my @nvidia story
January 2016. I'm an intern at a tech company nobody outside PC gaming cared about. Badge said NVIDIA. Friends asked if we made monitors lol
The whole place ran on email. Every team owed a weekly "top 5 things" - 5 bullets, no decks, no status meetings. It went up the chain and we were told Jensen had access to all of it and actually read them on Sunday nights.
Gave every intern this lowkey paranoia that the people way above us could see exactly what we were working on. They could.
I was too junior to send one myself, but some folks got replies right after midnight. Just 2 lines asking about bullet 4. The people who'd been there a while didn't even act like that was weird. That was the biggest tell.
Everyone there was ridiculously long $NVDA. My manager had an options spreadsheet he updated at lunch and refused to explain. Some people talked about the ticker like it was the weather.
Culture night had a whole skit about deep learning. Costumes, GPU jokes, "the network is learning," plus a bit about women supposedly going for guys who "know deep learning." Half the room laughed. Two people in that room were dead right about the next decade.
The part of 2016 nobody remembers: that exact same spring they were winding down Icera (the $367M modem bet) while posting record revenue. Record stock on one side of the building, a whole team getting walked out on the other. That's why "we're 30 days from going out of business" somehow never sounded like a gimmick. (He'd been saying it since '96 when it was literally true)
And then there's this week. Same CEO, same inbox, same voice. An email went to 42000 people who still carry that badge. An apology.. and now a resignation.
Also crazy to think if I'd thrown $3k into NVDA back then it'd be ~$600k today 💀 (even after this week's drop)
Anyway, Jensen replied to my mentor's bullet 4 at midnight and I'm keeping that flex forever
went down the rabbit hole testing openmausbot, paperclip, rakazo, and openbot.
ran them all on an always-on a macstudio m1
max at home, hitting it remotely over tailscale with heavy multi-agent / parallel subagent tasks.
ended up right back on hermes lol. nothing else actually holds up in daily use without constant babysitting
@NousResearch 🫡
@psipbc So we can never have exponential quantum speedups over unstructured data. Which means "Quantum AI" is dead (Or that term has a new meaning going forward).
If there’s one takeaway from the Binance hack, it’s this: going forward, attackers can seamlessly plan, code, social-engineer, and execute sophisticated campaigns using abliterated LLMs as their primary weapon.
Legacy playbooks won't work- we need new cybersecurity.