inimino

21 ms·

181.
by inimino 4y ago
The passionate people that started all these companies wouldn't take most jobs in them today. The bureaucrats have taken over, even surprisingly in "startup" culture, not just FAANG. I don't think programming will retain
182.
by inimino 4y ago
No, I don't think so, but DNS caching was mentioned, which is a somewhat similar optimization.
183.
by inimino 4y ago
Well, I think you're going beyond the parameters of the discussion... LLMs synthesize datasets and that is all they do. They are not reasoning agents and they don't have opinions about anything. All we can say is that they refle
184.
by inimino 4y ago
I think the point is when taking information from some source (like the web) to just represent that source fairly, which is a reasonable expectation of a search engine, for example.
185.
by inimino 4y ago
Good point.
186.
by inimino 4y ago
No, I agree. The Web is unbiased (as a reflection of the views of average web users, though that is itself a bias) or at least it mostly is (certainly was in the early days) and reflecting that evenly would be politically untenable. However
187.
by inimino 4y ago
Yes, it's a sample of all opinions in the training set. It has no opinion of its own, even no place of its own to stand from which to have an opinion. There can be no bias without reference to some ground truth, and there's no gen
188.
by inimino 4y ago
Close enough, I'm just peevish about that word. Teaching is miles away from anything in ML or deep learning practice today.
189.
by inimino 4y ago
"a growing number", "a lot". Where are those users now? We are a tiny lot, nearly economically inconsequential. Your prediction is optimistic.
190.
by inimino 4y ago
"Attention is all you need"? Joking aside, there's no reason AdWords can't become AdWordVectors and be even more effectively targeted.
191.
by inimino 4y ago
We're not even teaching it anything, all we're really doing is setting up a behaviorist training regime that lets it reproduce some of the biases of some of us, just well enough to squeeze through the Overton window that's ac
192.
by inimino 4y ago
I have always claimed that the best way to understand regexes is to implement them yourself. This goes for hashmaps and other basic data structures as well. In general this advice has been very poorly received. Going to the lower level (bui
193.
by inimino 4y ago
...and an IRC channel. The wiki wasn't much used, IRC was where everything happened. I think the way to do it today would be a discord server, and you need more people than you think to start; they will drop off fast.
194.
by inimino 4y ago
You're in good company, that's an extremely popular perspective. Those of use who are more generalist often have to find our own ways. The rewards are usually worth it.
195.
by inimino 4y ago
No, they taught it from 1980-97 with the original instructors, then for ten years the department head taught the introductory course (from unspecified materials, possibly different), then they went to Python with another set of new teachers
196.
by inimino 4y ago
It depends what you want to do. If you want to work on problems that we know how to solve and that we have solved many times, then the npm xenobiology skill set is sufficient; if you want to do things we haven't done and don't kno
197.
by inimino 4y ago
All you need is very rudimentary fluid dynamics, some vague idea of attack angle, and the conservation of momentum with a compressed column of air under the plane. I think most theoretical physicists could come up with that. The siphon one
198.
by inimino 4y ago
I believe engineering majors at any prestigious institution will be up to date on every one of the currently-known quarks, yes. If your point is that it's not the same as a physics major, that's also true.
199.
by inimino 4y ago
We both watched the same video. From my perspectice I see a subtext of the administration having its own (dysfunctional) process and getting a suboptimal outcome that's coming through very clearly. I mean, he admits the department stil
200.
by inimino 4y ago
You wouldn't have gotten into math the first time if it was just easy, we like it because it's hard.
201.
by inimino 4y ago
But do you choose to have that perception, or is it simply a consequence of looking at something that happens to be red?
202.
by inimino 4y ago
I kept my answer to about five minutes, so no to the transistors, although I started with a switch closing in the mouse or keyboard, so I was able to give some sense of a lower level.
203.
by inimino 4y ago
Believe it or not, both times were via HN. Someone set up a mailing list and wiki and we met every week.
204.
by inimino 4y ago
Well that's certainly a perspective, but you don't know what you don't know.
205.
by inimino 4y ago
You don't realize that Scheme used a garbage collected high level language at a time when C would have been the norm. This is like having 40000 npm dependencies today, convenient and very practical. The smart person was Knuth, who said
206.
by inimino 4y ago
I certainly do not own or want a self-driving car, at least not until they work! High school physics certainly does help you understand momentum, which is the cause of all crashes.
207.
by inimino 4y ago
Conducting a charity for the community is Mozilla's mission, so that part worked out. How good it will be for the community remains to be seen, at least it is probably replacing C++ on a lot of new projects. Nobody's saying you ha
208.
by inimino 4y ago
You can read some of the story of France's experiments with Bourbakism in elementary education here, under the section about pedagogy: https://en.wikipedia.org/wiki/Andr%C3%A9_Lichnerowicz tl;dr: New Math began in
209.
by inimino 4y ago
I'm editorializing a bit, his formulation of the idea is a little different from mine, but enough time in the industry makes it hard not to conclude that something is eroding the professionalism of the field, and I think his ideas abou
210.
by inimino 4y ago
There's a talk with slides that includes an overview of several of these projects: https://www.infoq.com/presentations/We-Really-Dont-Know-How-...

More ›