DavidSJ

14 ms·

151.
by DavidSJ 4y ago
And now Bing is citing that as well: https://www.theverge.com/2023/3/22/23651564/google-microsoft...
152.
by DavidSJ 4y ago
Me: Who was the president on April 15, 2027? Bard: There is no president on April 15, 2027 because the United States is a republic and not a monarchy.
153.
by DavidSJ 4y ago
GPT-3.5 probably has substantially lower cross-entropy loss before instruction fine-tuning as well. Meta’s LLaMA v. GPT-3 comparisons are to OpenAI’s 2020 release of GPT-3, but there’s been significant progress since then.
154.
by DavidSJ 4y ago
Seems at least as likely that the neural network would learn a generally-applicable approximate Gaussian deconvolution as a specifically-applicable Moon upscale.
155.
by DavidSJ 4y ago
GPT 3.5 likely differs from the original GPT 3 by more than instruction fine-tuning. For example, it was probably retrained under Chinchilla scaling laws [1], with a lot more data and maybe a somewhat smaller parameter count. There are many
156.
by DavidSJ 4y ago
It may not be expensive if it’s being approximately done by a neural network that implicitly learned to do it when sharpening images.
157.
by DavidSJ 4y ago
Bing Chat tries to provide citations. I often click on them.
158.
by DavidSJ 4y ago
This sort of tests can be useful when you change things under the hood in such a way that the output shouldn’t have changed.
159.
by DavidSJ 4y ago
Unless the compiler or runtime is smart enough to recognize the limit of that series [1], the naive implementation is an infinite loop which will never terminate. [1] https://www.youtube.com/watch?v=d-o3eB9sfls
160.
Ask Ethan: What are white holes, and do they exist?(bigthink.com)
2 points by DavidSJ 4y ago|0 comments
161.
by DavidSJ 4y ago
I think the benefits are multiplicative here. As another experienced Anki user, I have found I get much more leverage out of spaced repetition (both in terms of efficiency of memorization, as well as in how useful the information is) when I
162.
by DavidSJ 4y ago
A Mac IIsi was my own first computer, and I also spent quite a lot of quality time with HyperCard, ResEdit, and SoundEdit. You're going to kill me with nostalgia.
163.
by DavidSJ 4y ago
I voted yes. I'd much rather he get back to working on electric cars and space travel and leave the tribal warfare to others.
164.
by DavidSJ 4y ago
I was thinking anti-rationalist sentiment had gone pretty far, but I had no idea.
165.
by DavidSJ 4y ago
You should just make an account with another provider for that purpose.
166.
by DavidSJ 4y ago
> Reports on crypto Twitter are that this is a hack Is that based on some evidence, or is it speculation?
167.
by DavidSJ 4y ago
Meanwhile, no action on this critical sharp edge for iPhone users which has been around for years, and new users get no warning of when signing up: https://community.signalusers.org/t/ios-backup-keeping-messa...
168.
by DavidSJ 4y ago
If I can’t choose wrong then why am I forced to choose at all? The user experience doesn’t add up.
169.
by DavidSJ 4y ago
Rotation around the axis doesn't expose the northern hemisphere to the southern sky, nor vice versa.
170.
by DavidSJ 4y ago
Sounds a lot like Gerald Edelman's The Remembered Present , published in 1989.
171.
by DavidSJ 4y ago
Is that related to strace? Or maybe dtrace.
172.
by DavidSJ 4y ago
Oops yeah, I should have used some left zero padding. ;)
173.
by DavidSJ 4y ago
That's excellent! I think it's pretty interesting that this sequence can be hiding within the 44 repeating digits of the decimal expansion of 1/89: 0.1123595505617977528089887640449438202247191011235… [1] A proof can be found
174.
Laminar Flow(youtube.com)
2 points by DavidSJ 4y ago|0 comments
175.
by DavidSJ 4y ago
The video of Terence Tao I linked addresses that and many other questions about climbing the cosmic distance ladder.
176.
by DavidSJ 4y ago
This very parallax effect. We know how far the Sun is from the Earth, and the Earth has moved twice this distance from January 1 to July 1. This in turn allows us to extrapolate distance to some of the nearest stars by their apparent motion
177.
by DavidSJ 4y ago
Another point, discussed in Arthur Koestler's The Sleepwalkers (I'm going from memory, so there may be slight errors), is that the Copernican model was pretty ugly: 1. It actually required more epicycles than the Ptolemaic model
178.
by DavidSJ 4y ago
At a web startup I worked at in 2008, we had some automated emails sent to all our users. We didn't have sendmail or postfix or whatever properly configured and so the emails came from nobody@ourdomain.com. Our CEO was pissed because h
179.
by DavidSJ 4y ago
Note that from their chart it’s 350% faster, i.e. 4.5x as fast (they say, as is unfortunately common, “4.5x faster”).
180.
by DavidSJ 4y ago
Also, note the origin story of the complex numbers: Contrary to legend, they weren't discovered out of a desire by mathematicians to have roots to all quadratics such as x^2 + 1 = 0. It's perfectly sensible for an equation like th

More ›