Very tired
Chrysopolis
Joined January 2020
- Tweets15.4K
- Following374
- Followers635
- Likes27.9K
Can't quite come to terms with academics, scientists, specialists can be opportunistic grifters. Just doesn't fit together in my mind, so strange.
Shower thought: long Deleuzian century begins when applied statistics become robust enough to be tractable akin to a force of nature.
Saw like three people in public transport wearing both black and white Eintracht Frankfurt scarves and Raiders hats going to the match. Atlanticism is not dead.
I mean, the whole thing hinges on being vague whether it's an analogy or representation of human thought and what is human thought, so skepticism only takes us so far.
Ben Affleck says AI is simply matrix multiplication, linear algebra and floating point calculations, and however convincingly it sounds like a person, it has nothing to do with thought, humanity or feeling
“Nor do I believe, and I think it’s important to draw the distinction between what AI does, which is simply predict the next token or the next word in a sentence, then add that word on, run it back to the machine, do a lot of computation and pick the next word."
"That is a very different thing than sentience or thought. And I think we need to be really clear about that because it can be convincingly, in an illusory way, sound like thought or like a person."
"But let’s be real clear. This is matrix multiplication. This is linear algebra. These are floating point calculations. It has nothing to do with thought, humanity, or feeling.”
Bad news for all of us anti-aristoteleans since people are already interpreting this idea as proof that human language and neural networks interpret reality in a similar manner, as if any form of training didn't depend on human language including images.
interpretability might be the most important scientific problem in the world right now, as I don't think we can solve alignment if we don't know what we're building. I think I have good news about it. I co-authored early work in mechanistic interpretability at OpenAI and came away pessimistic, because neural networks appeared to fold into arbitrary tangled knots. years later, I started openai's first large-scale automated interpretability project, finding about 2% of neurons legible with the primitive tools we had
with modern agentic tools and a much closer look, the picture looks totally different. the vast majority in models up to trillions of parameters are legible, and their circuits beautiful. many famous examples of tangled neurons turned out to be misunderstood. The roadblocks I expected mostly aren't there. I now think ambitious interpretability will be solved by the time we need it most, and possibly quite soon, as automated scientists keep improving. it may be a bitter lesson for interpretability: the simple thing, at absurd scale, works
already-clean.goodfire.com/
More or less as expected, liberation from corporate overlords in editing software by means of other corporate overlords is postponed, because obviously it's a lot of work. Hence why I expect the impact on gaming industry to also be sharp but shortlived.
Ale×ey retweeted
How did Leibniz think on the page? Beyond polished works, hundreds of reading notes reveal a thinker testing possibilities, correcting errors, and reacting to his peers in real time.
More information about 2 forthcoming volumes of this material here: 🏛️ pfisogni.wixsite.com/rekh-ma…
One of the fun things about Richter is how much better his work looks on reproductions. From early monochromes to Baader Meinhof cycle, blurry images,monumental smudges, Polaroids - it's all too handmade in reality, but really shines on pictures. Another fun thing is his diaries.
Absolutely detest that with age knees and back develop their own whiny agency. How about we align that first
The naive thinking from a class that thinks politics are for stupid people, preferring to invent their own categories from first principles. We live in times of a political economy thats a kaleidoscope of nominalism for profit, opting out has no meaning in social sense.
Dario Amodei considers people opting out of AI-driven transformation of human life as a problem, perhaps a dystopian problem, needing to be "solved", on par with anti-vax. This is from the same essay, 'Machines of Loving Grace', in which he advocates that we become pets of the machines, and suggests video games as one of several alternate vectors of meaning to replace meaningful struggle and challenge.
