Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
superjan
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
11 ms
·
121.
▲
by
superjan
2y ago
Why work hard… Train a model to recognize the AI bots!
122.
▲
by
superjan
2y ago
And, refurbished business laptops tend to have better keyboards than consumer grade laptops, as well as a better build quality (generally speaking).
123.
▲
by
superjan
2y ago
Don’t worry. It _is_ a neat trick, and it was nice to post complete instructions and benchmarks. And it wasn’t you who put it to the HN frontpage.
124.
▲
by
superjan
2y ago
There’s a great paper from the group of Emery Berger, where they merge memory pages that become sparse. A kind of garbage collection for C and C++ languages. There’s also a good presentation about it on youtube. https://arxiv.org
125.
▲
by
superjan
2y ago
If you’re curious to try arthouse/international cinema, give Mubi a try. There is less to choose from, but the selection rotates.
126.
▲
by
superjan
2y ago
With respect to hallucinating, I never read about training LLM’s to say: “I don’t know” when they don’t know. Is that even researched?
127.
▲
by
superjan
2y ago
Does this also apply to “supersparpreis” tickets? i occasionally use DB but I find the pricing options and rules confusing.
128.
▲
by
superjan
2y ago
It’s a billion times integer modulus, summed. Div/mod is by far the slowest arithmetic instruction on a CPU, so it underestimates C/rust performance. I guess the author chose that to prevent the C compiler from optimizing the loop
129.
▲
by
superjan
2y ago
These extremes exist, and GR predictions are better than Newton’s in those cases. Closest to home is mercury’s perihelion drift. We have observed black hole mergers, gravitational lensing, and GR is also an essential component in understand
130.
▲
by
superjan
2y ago
The way I thought LCD/LED displays worked was by RGB filtering a uniform white backlight. Is it only this design that does fosforescence per subpixel? Sounds way more energy efficient.
131.
▲
by
superjan
2y ago
It depends on the application. It’s the easiest way especially if you might encounter right to left script, CJK or emoji. It is worthwhile to cache the textures, most text does not change every frame. It is good enough for us.
132.
▲
by
superjan
2y ago
I would argue it is conceivable that given enough time, a monkey or ape species can evolve that is intelligent enough to write Shakespeare’s full works.
133.
▲
by
superjan
2y ago
The RAM is expandable as well… however I am curious how well the extra RAM performs. Part of the M-series performance gain is from having the RAM dies very close to the processor.
134.
▲
by
superjan
2y ago
An example for those not in the know: to find a shortest route on a realworld map, an admissible heuristic would be that the minimum travel distance between two nodes will be a straight line. While examining options, A* takes this into acco
135.
▲
by
superjan
2y ago
As an alternative to simulating latency: How about using a VPN service to test your website via Australia? I suppose that when it easier to do, it is more likely that people will actually do this test.
136.
▲
by
superjan
2y ago
You’re right. I’ll read more carefully next time.
137.
▲
by
superjan
2y ago
The author wonders why he does not see a memory footprint for the billions of objects, and conjectures they are stored in-line in the array. I suspect that it is actually achieved by the flyweight pattern, where copies of the same immutable
138.
▲
by
superjan
2y ago
In such an alternative tech free universe, I guess most of us would be illiterate farmers.
139.
▲
by
superjan
2y ago
I have tried the windows version 2 years ago. The site I copied was our on-prem issue tracker (fogbugz) that we replaced. HTTrack did not work because of too much javascript rendering, and I could not figure out how to make it login. What
140.
▲
by
superjan
2y ago
I really love the LCD aesthetic of this site.
141.
▲
by
superjan
2y ago
It annoys me too, especially when the explanation is: “because RISC simpler so it is faster, duh”. However x64 is at disadvantage for instruction decoding (it’s difficult to decode multiple variable length instructions in parallel), and sec
142.
▲
by
superjan
2y ago
Such an experiment was described in a science book I read as a kid. They dissolved oil in alcohol so as to measure a very small quantity of oil, and then a a knitting needle was used to stretch the oil film across a plastic container filled
143.
▲
Building a 3D Printed Atomic-Resolution Scanning Tunneling Microscope (STM)
(youtube.com)
3 points
by
superjan
2y ago
|
1 comments
144.
▲
by
superjan
2y ago
How about applying the idea behind ESCROW: if you market hardware with software dependencies, you are required to provide the source to a trusted third party who will release/opensource it if you stop maintaining said software before t
145.
▲
by
superjan
2y ago
On a motorway, 50 to 80 eurocents per kWh is what I often see. There is not enough competition. It is still cheaper than what you’d spend on fossil fuels. However, for day to day use you charge at home or at the office which is much cheape
146.
▲
by
superjan
2y ago
Possibly they ask the text engine to render halfway two pixel lines. It will round up or down quite randomly when fp errors accumulate.
147.
▲
by
superjan
2y ago
I have recently been wondering how it is possible that they can make features like that, and provide closed captions, but that they continue to insert ads halfway a speaker’s sentence.
148.
▲
by
superjan
2y ago
Next level: let them login and forward the ssh connection to the digital equivalent of a room full of mirrors.
149.
▲
by
superjan
2y ago
The currently agreed protocol leads to < 10 transactions/sec. This can be improved, but 1000x will be challenging. https://en.wikipedia.org/wiki/Bitcoin_scalability_problem
150.
▲
by
superjan
2y ago
What I think is worth knowing is that compute units in GPU’s also use SMT, usually at a level of 7 to 10 threads per CU. This helps to hide latency.
More ›