jebarker

49 ms·

841.
by jebarker 3y ago
Most the AI doing image processing is architecturally very different and smaller than LLMs.
842.
by jebarker 3y ago
That word "purpose" terrifies me. Along with it's sibling "meaning". I (now) understand that purpose is something you define for yourself, i.e. it's not pre-ordained in nature for you. But I feel like the need
843.
by jebarker 3y ago
I'm 42 and I think it's relevant to me too
844.
by jebarker 3y ago
I really enjoyed this episode. The most surprising thing to me is how easily so many people seem to have been sucked in by the demo. I'm fairly sure I would never believe it was true until the clean-room test was shown to me. I guess I
845.
by jebarker 3y ago
You don't have to pay to rent the bike. The homeless person attaches the devices and then peddles on it stationary in the dock.
846.
by jebarker 3y ago
and median income in London appears to be ~41K GBP, so that rent would be about 80% of the income. No wonder there's lots of homelessness.
847.
by jebarker 3y ago
I don't think it's obvious that those traits, e.g. not stealing, are universal or part of some fundamental moral grammar. It just means they're common and widespread.
848.
by jebarker 3y ago
Ah, I see. So you can understand why another culture holds a particular moral position but you still _feel_ it to be immoral yourself. That makes sense. But how does that tie to the sparse data and acquisition of morality?
849.
by jebarker 3y ago
I'm not sure I understand the point being made. Chomsky seems to be saying that since we learn the prevalent morality of our culture through sparse data there must be an underlying universal morality. But how does that allow for the fa
850.
by jebarker 3y ago
I've watched a few episodes. As someone else pointed out, they choose locations where there's not going to be much available food, e.g. early winter Alaska. So you can't draw any conclusions about the viability of getting nut
851.
by jebarker 3y ago
I think that says more about the foraging skills of modern domesticated humans (and the location choices for Alone) than it does about the nutritional viability of a foraged plant based diet.
852.
by jebarker 3y ago
No because nature doesn't have intent.
853.
Upper Bound 2023: Insights into Intelligence, Keynote by Richard S. Sutton(youtube.com)
1 points by jebarker 3y ago|0 comments
854.
by jebarker 3y ago
Likewise. I've never found anything as effective as a pencil and paper as a scratchpad when thinking through problems. I also find a strange motivation day to day in checking items off a physical checklist/todo-list that I don&#x2
855.
by jebarker 3y ago
Did Terry Tao write that? EDIT: I see in a post further down that no he didn't. Ira Glass wrote it.
856.
by jebarker 3y ago
I definitely agree with perfectionism being used as an excuse. But how do people use "I was considered gifted as a child" to provide a similar excuse? Are they saying that was an incorrect assessment?
857.
by jebarker 3y ago
This isn't a color comment. It's a small correction to a factual error. No culture war to be had here.
858.
by jebarker 3y ago
Yes, your wording is better, I don't believe people are actively trying to be as wasteful as possible.
859.
by jebarker 3y ago
Nitpick: the author is female
860.
by jebarker 3y ago
I don't totally agree. It's what a short-term view of the economy rewards for sure. But even if that was the only view of the economy I've seen plenty of low-performance software written purely out of cargo culting and/o
861.
by jebarker 3y ago
This is true, but there's still choice in how we expand and the default seems to be to do it as wastefully as possible.
862.
by jebarker 3y ago
For sure, definitely not equating my PhD experience to Tao's. I felt like even in my viva the examiners didn't push me too hard and stuck close to the material in the thesis. Probably due to my decidedly non-prodigy status!
863.
by jebarker 3y ago
This is the first time I've read a detailed account of "generals" in a US math PhD. I'm surprised how broad the topics covered seem. I did my PhD in the UK and the transition to fully fledged PhD candidate after the firs
864.
by jebarker 3y ago
Nicely put. I'd also add that for me being a researcher became my identity and so my self-worth became tied to the perception and adoption of the research. With an established project (especially one I didn't write) there's a
865.
by jebarker 3y ago
I'm currently on a sort of sabbatical from greenfield research projects. I'm doing maintenance programming on a large widely used open source project. It's been really refreshing. I didn't realize how much I was feeling
866.
by jebarker 3y ago
Eh, I've worked with Math and Physics Ph.D.s throughout my career (and I am one myself). Those qualifications are definitely insufficient evidence on their own that you could beat the market. Not to say that hiring isn't their sec
867.
by jebarker 3y ago
I recently read Reanaissance co-founder Jim Simon's biography [1]. It seems like they discovered some secret algorithmic sauce that nobody else has found. But that feels implausible. Would love to understand what really gave them their
868.
by jebarker 3y ago
Would it provide useful/interesting results if the similarity search was global? E.g. find me neighborhoods in London most similar to this one in Chicago?
869.
by jebarker 3y ago
How well does (2) really measure accuracy? It seems like a single output that doesn't match the input code could indicate a fundamental floor in the optimized code, so it's essentially 100% wrong even though it gets the correct an
870.
by jebarker 3y ago
If there's a list of known optimizations that preserve correctness then it becomes an optimization problem based on output length (as a proxy for cycle count). So is the idea that an LLM is more efficient than a search or direct optimi

More ›