pradn

24 ms·

301.
by pradn 3y ago
The reduction in weight is a major feature - my iPhone 14 Pro routinely strains my hand.
302.
by pradn 3y ago
Nothing will change barring a change in economics (subsidies for quality games, Apple not taking a 30% cut on microtransactions, somehow prices for good indie games going up) or a change in human nature (people seem to just play shitty game
303.
by pradn 3y ago
Tensor libraries are high-level, so anything below them can be hyper-optimized. This includes the application model (do we still need processes for ML-serving/training tasks?), operating system (how can Linux be improved or bypassed?),
304.
by pradn 3y ago
I see, thank you! I hadn’t realized the hidden limitation in the usual iterator/generator/streams concepts in modern languages. It looks like the thing Linq and Java Streams have in common is that they’ve been added later, with a
305.
by pradn 3y ago
Oh I see, so with transducers, you can use the same transform pipeline (stack of functions) across different composition contexts (iterable next(), function composition, channel transforming, etc). I see the utility now! Still I think just
306.
by pradn 3y ago
I think these languages use “Collection” as another way of saying “iterable”. So the only thing such an object needs to provide is a “next()”. So a stack of lambdas/generators all working on a “next()” seem to be doing the same “per el
307.
by pradn 3y ago
I do wonder why people automatically assume "AI" is "100% AI"? Maybe we need to acclimatize. We all know when a company claims their soda is the tastiest or that their cars will make you cool, it's marketing fluff.
308.
by pradn 3y ago
This seems fine though? The postal office has workers who decipher the vanishingly small percent of mail automated systems can’t figure out.
309.
by pradn 3y ago
The difficulty is that in a VC world, you admit that you will more-or-less permanently need humans in the loop, which kills margins and scalability. At Google/Meta the number of SERP raters and content moderators are only a few tens of
310.
by pradn 3y ago
The problems with dimensionality certainly show up even with 256 dimensions. PCA-ing down to a few hundred dimensions is still a problem, and then you have to deal with PCA lossiness too!
311.
by pradn 3y ago
The elephant in the room is the Curse of Dimensionality. As the number of vector dimensions increase, "nearest neighbor" search behaves in an surprising manner. In a high-dimensional cube randomly sampled vectors are mostly near t
312.
by pradn 3y ago
A good chunk of the use of the word "creator" is the lack of a word to capture the work of someone who uses multiple platforms to distribute their work. Someone can be simply a YouTuber, but if they also do Twitch streams, they ha
313.
by pradn 3y ago
YouTube penalizes video creators who don't post regularly. So it's hard for them to take breaks. If a video takes a week to make, how will you have enough of a backlog built up to take a two week vacation?
314.
by pradn 3y ago
Non-use of enterprise data for training models is table-stakes for enterprise ML products. Google does the same thing, for example. They'll want to climb the compliance ladder to be considered in more highly-regulated industries. I don
315.
by pradn 3y ago
Most of the comments on this thread are talking about using OpenTelemetry to send metrics/logs to self-hosted collector jobs. Though using a standard library supported by a bunch of collector tools like ClickHouse is useful it itself,
316.
by pradn 3y ago
This is precisely the sort of article where a bit of notation and interactive graphics would go far in presenting the information. Anyone who can understand this article knows big-O notation. It's silly to assume the article is more ac
317.
by pradn 3y ago
It's hard for someone engaged in public policy to understand that most people just live their lives and have a basic and vague understanding of most policy issues. If you ask people about marine fishing limits, they'll say there a
318.
by pradn 3y ago
Late in the "enshittification cycle", competitors do rise, but they're usually under-resourced. So you get a overly-monetized primary product and a bunch of "never quite there" competitors. That's overall worse
319.
by pradn 3y ago
You should play Dragon Age: Origins. It's a lot tighter, way less filler. I look at these games with a million copy-paste quest markers and "points of interest" and just know there's little actual substance there. Or you
320.
by pradn 3y ago
1) Amdahl's law means it's not useful to have hundreds of cores for general purpose computing. There's not that much parallel work to do in typical applications. Increasing the proportion of work that's parallelizable fo
321.
by pradn 3y ago
If it was a joke, my bad !
322.
by pradn 3y ago
That's not similar at all! Windows takes the "Linux way" of "not breaking userspace" seriously. OpenBSD has other goals - some sense of minimality and clarity, but this is not one.
323.
by pradn 3y ago
I love how clearly this article is written. The author uses a table with the columns being "Initial Motivation for Fine-tuning" and "Why a Base LLM is Sufficient" - exactly what you need to learn why "You probably d
324.
by pradn 3y ago
Great post! Is there a public version of this chart? “ And that's what we're building right now: an alternate map of the world based not on geographic distance, but on pure network latency.”
325.
by pradn 3y ago
Re point #1 there has been a big shift in the corporate world to hiring vendors for any "non-essential" services. So now the chance of someone moving up from the mail room is low. There's been a progressive shift in how corpo
326.
by pradn 3y ago
The view outside of Hacker News on alignment ("censorship") is quite different. Two senators questioned Meta about its Llama 1 "leak". "While Meta originally only purported to release the program to approved researc
327.
by pradn 3y ago
There is a post about their font design actually. It's by Lettermatic, the font design house that helped them with the game. https://lettermatic.com/custom/pentiment
328.
by pradn 3y ago
I mean, maybe. AI on the "good side" will also improve. It should be possible to check a sentence against its reference with LLMs. And anything not sourced is suspect, just as it is now. I also don't like the attitude of Wiki
329.
by pradn 3y ago
I'm unsure if this will happen. There's plenty of checks-and-balances for Wikipedia edits. There's automated spam detection, editors manually looking over edits for articles on their watchlist, editors who look over subtopics
330.
by pradn 3y ago
The US public seems to have the capacity to absorb an incredible amount of misery while just acting up at tame things like elections. There are occasional protests and riots, but really quite few - and they dissipate quickly. Hell, we had 3

More ›