Jasmine Li

21 followers · 19 following · 1111 views

on the atlas — 36

highlights — 24

Everything in life deserves a full ass.
2024 year in review | Kevin Liu
You can find my most up to date thoughts on artificial intelligence in my annual letter: 2025, 2024, 2023, 2022. I also recommend the following to any researcher, in order of increasing commitment to consume: Jorge Luis Borges, On Exactitude in Science Laura Deming, The Rage of Research Kanjun Qiu, Research as Understanding David Pfau, On Slow Research Michael Griffin, Space Exploration: Real Reasons and Acceptable Reasons Nelson Elhage, Computers can be understood C. S. Lewis, Learning in Wartime SLIME MOLD TIME MOLD, The Scientific Virtues Nima Arkani-Hamed, The Morality of Fundamental Physi…
Readme | Zhengdong
So. Two alternatives: Focus on your marginal impact, instead of your absolute impact. No one person’s marginal contributions, in expectation, are going to be able to swing p(doom) by 1%. A much more reasonable target to aim for on a personal level is .01% or .0001% or so. Or: the paths to successful worlds are highly irregular. There might be several different lines that will get us there, many requiring a high number of steps in sequence, an unknown number of which are interchangeable. The problem is too difficult and unknowable to model with a single final probability, or is simply not even …
Turning 20 in the probable pre-apocalypse — LessWrong
I promised I would give concrete advice -- so let me share with you an actual rubric I use when choosing research projects. This is slightly off-theme for the rest of the talk. It is partially informed by AGI -- but is a good set of questions to ask about a project regardless of AGI timelines. First, you should gate by impact. If this project works flawlessly, how large is the potential benefit? Make sure you evaluate the expected outcome projected onto your value axis, rather than just measuring its norm. See the previous section -- you want to have an impact you can be proud of, and this mat…
Jascha Sohl-Dickstein on X: "Title: Advice for a young investigator in the first and last days of the Anthropocene Abstract: Within just a few years, it is likely that we will create AI systems that outperform the best humans on all intellectual tasks. This will have implications for your research and https://t.co/vnST8BbKdH" / X
"But doesn't everything always just work out? Like sure, it might suck in the moment, but everything's always fine. Somehow every moment in life always magically works out in the long run. I don't think I could name one time where everything went completely wrong, because if it did I wouldn't be here right now."
Training the Idea Muscle | Alexandria
Riley's biggest inspiration is this guy who makes $200k/year writing music that you've probably never heard of: Matt Farley. Matt's been making about 50 songs a day for the past 12 years, but most of those songs barely get any streams. He's written a book, produced short films, all of which have been seen by maybe a few hundred people total. But when you add up the impressions of his entire 20,000-song body of work, it starts to make sense why he makes that much money. When asked about how he makes so much stuff, he said: "If you reject your own ideas, then the part of the brain that comes up …
Training the Idea Muscle | Alexandria
Borderline consciousness, when it’s neither determinately true nor determinately false that experience is present
Borderline consciousness, when it’s neither determinately true nor determinately false that experience is present | Philosophical Studies
Zir children have never known what it’s like to be at war within themselves, and hopefully never will
The Gentle Romance - by Richard Ngo - Asimov Press
tending their mental gardens since infancy.
The Gentle Romance - by Richard Ngo - Asimov Press
Over the next decade, his meta-self grows even vaster, taking up hundreds of GPUs, with his biological brain just one small component of it. Elena’s grows in synchrony, with well-worn connections between them where they send thoughts directly to each other’s minds. Learning to be so open with each other isn’t easy, though. He’s ashamed to let Elena see how lost he’d been before her. And she worries that if he understands how intensely she fears abandonment, it’ll become self-fulfilling. Working through these fears strengthens their trust in each other, allowing their minds to intertwine like t…
The Gentle Romance - by Richard Ngo - Asimov Press
His consciousness lingers on the most novel and informative conversations, while copies of different parts of his mind survey all the information he receives in detail.
The Gentle Romance - by Richard Ngo - Asimov Press
He calls the voice his meta-self; as it learns to mimic him more faithfully, he increasingly comes to rely on it. He can send his meta-self into meetings with someone else’s meta-self, and they’ll often be able to make decisions or delegate responsibilities without bothering him. His meta-self helps him navigate outside work too. He’s now a regular at the book club, but he hasn’t had much practice at making friends and sometimes feels out of place. He recruits his meta-self to tell him when he’s doing something rude, and to talk to his friends’ meta-selves to figure out how to defuse any confl…
The Gentle Romance - by Richard Ngo - Asimov Press
How well do you expect to do if you continue the current trend? What would it be like to do better than that?
Tune Your Cognitive Strategies
I want to emphasize here is that compression is effortful
When is better to think without words? - by Henrik Karlsson
Confronted by a vast array of books, anyone will be seized by the anguish of learning, and will inevitably lapse into asking the question that expresses his torment and his remorse.
Inbox (2,175) - anishk@berkeley.edu - UC Berkeley Mail
“Because when I read, I don’t really read; I pop a beautiful sentence into my mouth and suck it like a fruit drop, or I sip it like a liqueur until the thought dissolves in me like alcohol, infusing brain and heart and coursing on through the veins to the root of each blood vessel.”
Text remains the loveliest medium - Tommy Dixon
Humans have really diverse patterns of thought.
What I Think About When I Think About History — LessWrong
influential METR AI research center
AI’s capabilities may be exaggerated by flawed tests, study says
we can generate scientifically promising ‘alien’ hypotheses unlikely to be imagined or pursued without intervention until the distant future, which hold promise to punctuate scientific advance beyond questions currently pursued
Accelerating science with human-aware artificial intelligence | Nature Human Behaviour
That, of course, is easier said than done. Perhaps the most modern tools we have available – namely, powerful AI systems – could help us. It is possible to train an AI to escape the disciplinary lines of universities, instead generating ‘alien’, yet scientifically plausible, hypotheses from across the entire scientific literature. These might be based, for example, on the identification of unstudied pairs of scientific concepts, unlikely to be imagined by human scientists in the near future. It’s already been shown in research on natural language processing that a purely textual analysis of pu…
Waking up science’s sleeping beauties - Works in Progress
Here I compare the accuracy of my annual predictions from the past 5 years (2020-24) with the previous 6 years (2014-19). The chart shows for each confidence level (x-axis) how many predictions at that confidence level were correct (y-axis), e.g. “out of all the predictions I made with 70% confidence, how many of them came true?”. The green line of “perfect calibration” shows the ideal level of correctness for reference, e.g. “predictions made with 70% confidence should come true 70% of the time”. Compared to 2014-2019, my overall calibration seems to have become worse (the correctness line is…
2024-25 New Year review | Victoria Krakovna
Whatever you’re working at today, you might consider the possibility that it’s also pre-competitive, even if it feels “late” to other people. Many, many human beings had engaged in distance running before the four-minute mile was achieved in 1954. So, if a given skill plays a crucial load-bearing function in your life, try to completely ignore the standards others set and assume there’s a much higher standard to which you can aspire. Try to internalize this insane belief: “I’m going to be better at this than anybody I’ve ever seen.”
Gmail
It keeps becoming increasingly obvious that we can make massive Pareto improvements over classical school. This is the most glaring example. The only big disadvantage that actually matters is that it remains expensive, but that will improve over time, and for what you get it is already a fantastic deal.
Gmail
bliss as an object of scientific study
The Meditation Start-Up That’s Selling Bliss on Demand - The Atlantic