Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
porphyra
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
16 ms
·
301.
▲
by
porphyra
1y ago
Star Trek is often considered the archetype of optimistic science fiction.
302.
▲
by
porphyra
1y ago
The iterator in my example will also happily iterate on things that don't exist. We all agree that it's possible to do in any language. But the main point is that the blog post is talking about syntactic sugar for an easier way to
303.
▲
by
porphyra
1y ago
Well, the whole point of the blog post is to argue for a new kind of syntactic sugar and the author explicitly mentions iterators. > Well a range based for loop requires that your tree exist in memory AND that you have an iterator define
304.
▲
by
porphyra
1y ago
The operating speed on the fast stretch is actually one of the fastest in the world at 350 km/h, it's just that it has to waste half an hour on the Caltrain tracks from San Francisco to Gilroy and from Burbank to Anaheim. It'
305.
▲
by
porphyra
1y ago
Other countries can build entire rail lines entirely on viaducts for the same cost as it takes to grade separate a single Caltrain station. For example it was recently estimated that grade separating Broadway Station with a center platform
306.
▲
by
porphyra
1y ago
Good for them! These kinds of student projects are very helpful and hopefully the students will continue working on these cool things when they graduate.
307.
▲
by
porphyra
2y ago
It's not about "inflating" it. It's more that the energy needed to move your car a certain distance is quadratically related to the speed, due to aerodynamic drag. Efficient vehicles spend less energy on other stuff besi
308.
▲
by
porphyra
2y ago
At $20k it is actually comparable in cost to a GEM el Xd pickup [1] which can only go up to 35 mph, has 78 mile range, and costs $18k [2]. Totally different class of vehicle, of course. [1] https://www.gemcar.com/gem-el-xd&#
309.
▲
by
porphyra
2y ago
Isn't the ip rotator used to scrape from public websites to bypass rate limits? Not sure how that automatically means they are "siphoning sensitive case files".
310.
▲
by
porphyra
2y ago
It is a technological tour de force and an amazing demo of what's possible and what is soon to come. But if we define a "good product" to mean a commercially successful one, then it isn't very successful. Still, I'm
311.
▲
by
porphyra
2y ago
I personally think that it is amazing that we are likely going to shift away from fossil fuels out of pure capitalistic greed. Trying to convince people to give up fossil fuels out of some altruistic desire to save the planet was never gonn
312.
▲
by
porphyra
2y ago
At least the percentage of renewable energy generation in China is steadily increasing. It has gone from 17% back in 2008 to 32% in 2024. And solar and wind are growing at a rapid pace. [1] https://en.wikipedia.org/wiki/
313.
▲
by
porphyra
2y ago
It is funny that Microsoft had been peddling "AI PCs" and Apple had been peddling "made for Apple Intelligence" for a while now, when in fact usable models for consumer GPUs are only barely starting to be a thing on extr
314.
▲
by
porphyra
2y ago
Animated webp has pretty good browser support by now and Discord uses it by default to serve animated emojis and stickers. However, many image hosting tools still don't let you upload webp. [1] https://caniuse.com/webp
315.
▲
by
porphyra
2y ago
True, but embedding a gif is way easier than using a javascript thing which might not be allowed in most places.
316.
▲
by
porphyra
2y ago
To be fair, terminal output is one of the few things where GIF's LZW compression and limited color palettes shine at.
317.
▲
by
porphyra
2y ago
Yeah most of the information is lost, but if you know the font used for the text (as is the case with a screencast of a macOS window), then you can try every possible combination of characters, render it, and then apply the averaging, and s
318.
▲
by
porphyra
2y ago
OpenAI themselves said > One last note: we’ll also begin deprecating GPT-4.5 Preview in the API today as GPT-4.1 offers improved or similar performance on many key capabilities at lower latency and cost. GPT-4.5 in the API will be turned
319.
▲
by
porphyra
2y ago
So, whenever you fail to change someone's mind, you can just dismiss them as being a toddler. This mindset explains how the current state of, say, US politics became so polarized and extremist.
320.
▲
by
porphyra
2y ago
AMD GPUs actually have fairly powerful hardware. However, despite being open source, calling them "developer friendly" is rather generous. Just look at how TinyCorp struggled with AMD ROCm before resorting to writing a completely
321.
▲
by
porphyra
2y ago
Merit-based admission sounds good to me. Harvard is vigorously defending its "right" to continue to deny admissions to highly qualified Asian applicants out of nothing but pure racism, and somehow they are the good guys?
322.
▲
by
porphyra
2y ago
pretty wild versioning that GPT 4.1 is newer and better in many regards than GPT 4.5.
323.
▲
by
porphyra
2y ago
I could even see Gemini 2.5 googling the right things and "thinking" about Gerald Ratner before it abruptly censored itself at the last moment.
324.
▲
by
porphyra
2y ago
As long as Google continues to hamstring themselves with censorship for no reason, I can't use their products. The other day I asked gemini 2.5 pro "which british ceo said that his company's products were bad" and the re
325.
▲
by
porphyra
2y ago
How can they tell if you submit a patch that you produced with the help of AI though? They mentioned they found a bug in an AI-assisted patch. But an actual skilled programmer can easily produce code with the help of AI that doesn't ha
326.
▲
by
porphyra
2y ago
Very cool to see this project get rebooted. I'm hoping it will have the critical mass needed to actually take off. Writing CUDA kernels in C++ is a pain. In theory, since the NVVM IR is based on LLVM IR, rust in CUDA should be quite do
327.
▲
by
porphyra
2y ago
With GCC error messages usually the only part I wanna see is "required from here" and yet it spams zillions of "note: ...". With the nesting enabled, it's way easier to read --- just look at the top indentation leve
328.
▲
by
porphyra
2y ago
Chatgpt 4o's advanced image generation seems to have a low-resolution autoregressive part that generates tokens directly, and an image upscaling decoding step that turns the (perhaps 100 px wide) token-image into the actual 1024 px wid
329.
▲
by
porphyra
2y ago
I think that one reason that humans are so good at understanding images is that our eyes see video rather than still images. Video lets us see "cause and effect" by seeing what happens after something. It also allows us to grasp t
330.
▲
by
porphyra
2y ago
But with modern hardware, running the dynamic programming solution to this optimization problem takes a trivial amount of cycles* compared to rendering your typical React webapp. * for most webpages. Of course you can come up with giant ebo
More ›