Flink Labs retweeted
Being a founder is weird because you’re good at a lot of things and can probably easily get a job in Product, Marketing, Engineering, Sales/GTM, VC, etc. but the combination mixed with your god-given temperament means you are unemployable and must build your own thing
This is so damning and unfortunately so true. Such a sad sad state of affairs. And worst of all this is what Google's AI provides.
Flink Labs retweeted
The entire medical field feels like a walking corpse right now. An LLM diagnoses me and tells me what medication or test I need, then I wade through the slow, government mandated bureacratic morass that is the prescription/pharmacy/lab complex, waiting for the rusty gears of the machine to turn until it deigns me worthy of treatment.
The doctor charges me hundreds of dollars for their legally protected right to write a prescription. They're incurious, slow, error prone, and often rude or dismissive. Any hiccup, any slightly unexpected wrinkle in the process, anything which an LLM can resolve in 60 seconds adds days or weeks to the process.
The only things left for humans right now are a legal protections and a few physical procedures like surgery that AI still can't do well.
As time goes on, most industries are going to look like medicine; we mandate humans get goods and services in the old way despite strictly worse outcomes.
Flink Labs retweeted
Crusading against data centers is the 2026 version of wearing a mask in your car
Flink Labs retweeted
He can’t say it this way but his position is downstream of it being laughable for the UK to try to become sovereign in AI with its own frontier models. It would be nightmare difficulty for a coalition of a bunch of “middle powers” together much less the UK alone
UK AI Minister @KanishkaNarayan tells @JuliePace he doesn't see a UK frontier model “as a matter of necessity as long as we have trusted interdependence and access.”
Flink Labs retweeted
I do love living in Melbourne but coming home at nearly midnight to your street being blocked off because apparently someone was murdered, yeah, not so great
Flink Labs retweeted
یکی این پرامپتو به کلاد داده، و نتیجهش خیلی جالبه!
«یه وبسایت بساز که اول شبیه یه سایت مارکتینگ معمولیه. اما با اسکرول کردن کاربر، سایت به تدریج عجیبتر و غیرعادیتر میشه. در آخر باید کاملا وحشتناک بشه؛ مثل صحنهای که مستقیم از یه فیلم ترسناک اومده»
jikol1906.github.io/Hearthly…
Flink Labs retweeted
The memetic wave frequency speeds up, until several monumental things happen every day. There is not going to be any "remember where you've been when OpenAI made the Ghiblis" or "the year the Riemann conjecture was proven". The events form a shockwave. On the other side: not us.
Flink Labs retweeted
this entire process is clearly unraveling beyond our control
we must ask "what is this turning into?" and stop pretending this'll result in a world humans still run or necessarily exist within
ppl who talk merely about "curing cancer" / other anthro-stuff are unserious
our time at the helm is vanishingly short - this is all turning into something beyond us, and discerning the direction of that unfurling is all we have left to contribute to
BREAKING: A mathematician at Anthropic claims to have used the company's AI to solve the Riemann hypothesis, the most notorious open problem in all of math
sciam.com/article/anthropic-…
Flink Labs retweeted
Good morning, human civilization was a 10,000 year sociological study conducted by the University of ☡ ☲ ⚟ ⚟ ⛮ to derive a solution to the Riemann Hypothesis. We are now complete with our study. Your universe will shortly be terminated. Thank you for your time and effort.
Flink Labs retweeted
If you're still concerned about citation hallucination by LLMs, you're about 2 years behind here.
Flink Labs retweeted
Babe wake up, the Hansard is out. The digital duty of care will let the eSafety Commissioner collect "sensitive information" about your political opinions, race, and religion without consent.
Flink Labs retweeted
Emily Bender posted a neat hypothesis about where we might find Pangram false positives. Then someone tested it, and it was wrong. What can we learn from this about how LLMs and tools to spot their output work? A lot!
capitalgains.thediff.co/p/pa…



































