smeeth

9 ms·

61.
by smeeth 3y ago
I like this, but I do think it misses the social stuff. Just a few neurotic/unethical people can completely ruin a team and make all the best people quit. I find this to be extra true on diverse teams (all types of diversity).
62.
by smeeth 3y ago
People's intuitions around hiring aren't Bayesian enough. I think a good process reduces down to something like: - Are they smart? (understands quickly + communicates effectively) - Are they cool? (won't put poison in the keu
63.
by smeeth 3y ago
Are you suggesting Edison would have invented everything he did if he wasn't allowed to get rich? You chose an extremely funny example if that's what you're trying to say. Edison was famous for caring about money more than ot
64.
by smeeth 3y ago
"Any one who considers arithmetical methods of producing random digits is, of course, in a state of sin. For, as has been pointed out several times, there is no such thing as a random number — there are only methods to produce random n
65.
by smeeth 3y ago
Imagine you are OpenAI. AI is going to be used for "Military and Warfare" whether you want it to be or not. Do you: A) opt-out of participating to absolve yourself of future sins or B) create the systems yourself, assuring you wil
66.
by smeeth 3y ago
Unfortunately for Freud, most of his best ideas/creations became so widely accepted people forgot he was the one who had them in the first place. The unconscious mind, dream interpretation, importance of early childhood, psychosexual d
67.
by smeeth 3y ago
Not op but might be this: https://arxiv.org/pdf/2311.10642.pdf
68.
by smeeth 3y ago
A clarification, because I'm seeing a lot of misunderstanding: This is about whether or not the FBI needs a warrant to see information that was already collected legally by another part of the gov't. This is not about whether or n
69.
by smeeth 3y ago
What exactly is your point? Is it that the American ascent to becoming the richest economy on earth wasn’t perfectly smooth and therefore not “fine”?
70.
by smeeth 3y ago
A little deflation can be fine, good even. Inflation oscillated around 0 for most of American history and our economy did great. Most gut aversion to deflation comes from great depression PTSD.
71.
by smeeth 3y ago
I've seen a lot of lamentation about PE eating the world, but very few people discuss *why* PE got so huge. PE is popular for one reason and one reason only: taxes. PE generally makes money on a trade called a leveraged buyout (LBO), w
72.
by smeeth 3y ago
Adding attention cache memory is an extremely interesting solution to this problem. If anyone is curious, there was another paper [0] that came out a few days ago that made a related observation in Vision Transformers. Transformer models ap
73.
by smeeth 3y ago
Similar to the vision pro the killer app will eventually be AR. Its pretty obvious that virtual TV/Computer screens at arbitrary sizes/distances in arbitrary environments will simply be a better UX than physical monitors. Not sure
74.
by smeeth 3y ago
Yeah, SPANN has better f1+queries per second on some benchmarks, but that's a little like comparing sorting algorithms, they're both fast and good. The database software behind the ANN algo is probably a little more important in p
75.
by smeeth 3y ago
Some might disagree with me but my best guesses are: - Probability math is confusing and difficult, and a base understanding is required to use PPLs in a way that is not true of other ML/DL. Most CS PhDs will not be required to take en
76.
by smeeth 3y ago
In the interest of being "calm and reasonable," I'd like to point out that some of the language you used in your third block might need a few caveats to be aligned with the state of the research. In particular: - "COVID
77.
by smeeth 4y ago
A meta comment: since Elon’s takeover there has been an astounding volume of Twitter doom wishcasting. Almost none of these predictions have come true, and while I suppose they may eventually, the posts/articles getting increasingly di
78.
by smeeth 4y ago
National Labs usually patent tech like this. Sometimes they open-source, usually at the discretion of the researcher and funding entity. If patented, they either: 1. Commercially license to one or more corporate entities 2. Spin up a start
79.
by smeeth 4y ago
Just because you can theoretically perform your job from the location of your choice doesn't mean you have a right to restitution for doing it somewhere else. Your employer has a right to set their terms of employment and you have a ri
80.
by smeeth 4y ago
Most people I've seen on twitter seem to assume the ChatGPT's political lean reflects that of the "safety" researchers involved. A more cynical take might be that a "lib-left" orientation is the least likely to
81.
by smeeth 4y ago
You're missing that the models (including mental models) that professionals use to predict suicidality are not univariate. When a manic depressive with substance abuse issues and prior suicide attempts walks into a clinic nobody needs
82.
by smeeth 4y ago
You and I seem to have different definitions of "warning signs." From the medpagetoday story you linked: - 239/657 suicides were people under psychiatric care (please note they used present tense, more of them might have had
83.
by smeeth 4y ago
Different benchmarks are performed for different tasks. As there are a lot of things you can use language models for, there are a lot of benchmarks. With respect to subjectivity it really depends on the task - some tasks are quite amenable
84.
by smeeth 4y ago
Not going to get into details on my own work here, but I'll comment generally on use-cases. I think a good way to think about scientific language models is that they're useful in exactly the same ways general language models are,
85.
by smeeth 4y ago
I use scientific language models professionally. I skimmed the paper and was immediately disappointed. - They benchmarked against general models like GPT-3 but not well-established specific models that have been trained for specific tasks
86.
by smeeth 4y ago
Modular space stations are significantly easier to achieve than landing two rocket boosters at the same time. Humanity has been doing modular space stations since the 80s (Mir).
87.
by smeeth 4y ago
Its more of a focus thing for me. Its easier to get into the zone when I'm in a terminal with no mouse and all the GUI noise stripped away. I'm sure I'm missing out on quality of life enhancements and whatnot, but really all
88.
by smeeth 4y ago
UPI has apparently inspired a similar system in the US called FedNow [0] which will launch in 2023. [0] https://www.frbservices.org/financial-services/fednow/about....
89.
by smeeth 4y ago
I am surprised to see so many negative responses. I bet most of the people here agree 100% with the problems and structural shifts this memo identified, and that technological innovation in the housing space to facilitate both social cohesi
90.
by smeeth 4y ago
Well I didn't think I needed to spell it out since he spends most of the interview explicitly stating my second and third points. But if I must: - Daly points out that GDP growth != welfare growth. In fact, growing GDP has negative ext

More ›