Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
toxik
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
12 ms
·
181.
▲
by
toxik
2y ago
I know we’re not supposed to whine about the medium but focus on the content. However, on Medium, the medium becomes the content because I can’t even read the damn thing without being accosted by some CTA popover. This is my quiet crusade a
182.
▲
by
toxik
2y ago
Oh I didn't mean to sound dismissive, it's just easier to understand it as a toolkit defined mostly by necessity.
183.
▲
by
toxik
2y ago
It is a by-product of a research project, its main connection is "these things were useful to the author while working on spectral images".
184.
▲
by
toxik
2y ago
I feel like ”quite well” is overselling it a bit. It did maybe better.
185.
▲
by
toxik
2y ago
This is generating SVG data, not using an image generator.
186.
▲
by
toxik
2y ago
Sounds like Simulink to me, which is one of the major MATLAB programming environments. ETA: Apparently MathWorks has Simscape in this category.
187.
▲
by
toxik
2y ago
Ah, it is the one in economy that is not actually an official Nobel prize, but a prize in his memory or something like that. https://en.m.wikipedia.org/wiki/Nobel_Memorial_Prize_in_Econ...
188.
▲
by
toxik
2y ago
Most priors are not encodable as architecture though, or only partially.
189.
▲
by
toxik
2y ago
The peace prize isn’t awarded by the Nobel committee either.
190.
▲
by
toxik
2y ago
Do you know what a Y combinator is?
191.
▲
by
toxik
2y ago
These CEOs have found a way to game the system, but the system presupposes mutual respect. If one side reneges on this contract and says that technically they're following the rules, well, the other side will just disregard the rules.
192.
▲
by
toxik
2y ago
1. not annoy no one 2. neg (annoy no one) 3. neg (neg exist x : annoy(x)) 4. exist x : annoy(x) 5. annoy someone
193.
▲
by
toxik
2y ago
Hear, hear. I often spend five times as long peddling about with the type annotations. Most of the “bugs” I find with the type checker are type annotation bugs, not actual software bugs. What type annotations do however deliver is useful co
194.
▲
by
toxik
2y ago
Predicting next moves of some expert chess policy is just imitation learning, a well-studied proposal. You can add return-to-go to let the network try to learn what kinds of moves are made in good vs bad games, which would be an offline RL
195.
▲
by
toxik
2y ago
Why do you think that? InstructGPT was predominantly trained as a next-token predictor on whatever soup of data OpenAI curated at the time. The alignment signal (both RL part and the supervised prompt/answer pairs) are a tiny bit of th
196.
▲
by
toxik
2y ago
Meta comment here, but Science Direct is an aggregator, and it doesn't make sense to talk about it as publishing. Elsevier published the referenced work in the journal "Advances in Nutrition", vol 8, issue 3, pp. 511-519.
197.
▲
by
toxik
2y ago
I was so disappointed that this turned from Python to Haskell, I would just _love_ to have something like this in Python. Even just type annotations. A similar thing I have been thinking about doing in robotics contexts is to annotate vecto
198.
▲
by
toxik
2y ago
What is presented here is a facsimile of interaction, and it is sorely lacking while being presented otherwise. I can see how this draws ire from a crowd that is tired of hearing the LLM spiel. For reference, I input "I solve the case&
199.
▲
by
toxik
2y ago
Hierarchy aside, I am surprised the literal author and maintainer of the project, on Google’s payroll no less, was not consulted on such a decision. Seems borderline arrogant.
200.
▲
by
toxik
2y ago
From that paper it seems the sampling method (SAP) is also slower, so that it beats larger models seems expected.
201.
▲
by
toxik
2y ago
Unitree products feel like oversized toys, not serious research platforms.
202.
▲
by
toxik
2y ago
I think it’s okay that they use the contract force to remove backlash since they are actually controlling it. In fact, if you could do that well, that’s huge!
203.
▲
by
toxik
2y ago
I don’t understand why you think they are even similar. This is still doing pairwise attention.
204.
▲
by
toxik
2y ago
I write Dvorak every day and I have no problems using QWERTY. I still type it faster than most QWERTY only typists, and with Dvorak I am basically undefeated.
205.
▲
by
toxik
2y ago
A lot of tools strip metadata by accident or intent, I have come to find out. Uploading to Slack, for example, strips ITXT from PNG files.
206.
▲
by
toxik
2y ago
I think it’s helpful because it is a common mistake.
207.
▲
by
toxik
2y ago
No, and it’s absurdly petty to argue otherwise.
208.
▲
by
toxik
2y ago
O(microseconds) = O(years), this is not what big O notation means.
209.
▲
by
toxik
2y ago
Measure waist circumference over time.
210.
▲
by
toxik
2y ago
If an individual's environment causes them to self-harm by stabbing themselves with a knife, you take away the knife before (if) you look at the root cause. The body wants to maintain whatever weight it usually has, if you can keep tha
More ›