@nntaleb The same fraud plays out in healthcare all the time.
“Evidence-based” often just means selectively measured, poorly understood, and heavily incentivized.
First principles thinking rarely survives systems addicted to protocol.
@kimmonismus Because he is half the age of his “peers”
Because he does not get lots of $$$ hush money to conform (he is not on the radar of the big companies)
He is not in for the career
He is just not corrupted :(
What is the best part of Databricks?
For me, Unity Catalog is impressive.
Quite a boring lookup table but it connects so many things and gets the job done.
Great governance is what makes a data lake useful.
@FlorianGallwitz Wird natürlich nicht passieren, Europa ist ein Museum.
Das mit dem Rückstand sehe ich aber überhaupt nicht. Die meiste Forschung ist offen zugänglich. Mit einem entsprechenden Budget bekommt man key Personen aus den Top Labs. Maximal 6 Monate Rückstand
Teaching is considered the gold standard of learning (Feynman technique)
Learn by simplifying a concept - identify & eliminate the unnecessary - get feedback on your understanding and explanation.
Writing forces structure. Give feedback to yourself. Your thinking will improve.
Nonsense like this is the best reason for UBI. It’s so sad that a lot of STEM is rather stagnant. But if the best minds think about career and grand money instead of the science who wonders.
Great video, now what ? :( https://t.co/iHeSGltWXD
@tegmark Put the paddle to the Metal. All these regulations will not help with anything but make sure there is no AI in Europe but imported from US/China
What did the EU founders of Mistral and Black Forest Labs have to say about EU AI regulation?
Beautiful podcast by @OfficialLoganK great points about why Open Source models are king (they are more controllable and therefore can be treated as more deterministic)
Also the points about AI in large corp vs startups made me think a lot. https://t.co/GBuU0Kw7BE
@NoahCarl90@plk669888 JD also directly addressed that: ppl who push for AI safety regulations and slowing down are mostly doing so for regulatory capture.
We’ve found as AIs get smarter, they develop their own coherent value systems.
For example they value lives in Pakistan > India > China > US
These are not just random biases, but internally consistent values that shape their behavior, with many implications for AI alignment. 🧵
@DaveShapi@packyM Literally this
Nukes are so much worse then everything else, yet here we are.
Have some faith in humanity (especially if all of humanity can level up their IQ with AI)
@vonderleyen I actually agree that it’s not to late. But we need really change course. JD Vance speach was spot on. Europe needs to be smart and fast now. Break things in the process.
We need free access to high quality data in every domain. Don’t regulate but demand that!
Some people, misled by their school teachers, try to understand things by ingesting narratives, with the goal of reproducing the narrative. For them, discovering truth is about whom to believe. That's futile. Understanding requires grappling with reality. https://t.co/ikbPNXcPsQ