Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
deckar01
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
13 ms
·
181.
▲
by
deckar01
2y ago
I don’t understand the use case. You go through the trouble of generating checksums when copying videos, but don’t want to regenerate the checksums when modifying the metadata? If you are this concerned about data corruption why not check t
182.
▲
by
deckar01
2y ago
There is good open source cutter plotting software available, but Cricut sues everyone who bypasses their app. I suspect their hardware is being subsidized by these asinine paywalls to monopolize the consumer cutter market.
183.
▲
by
deckar01
3y ago
What I need is a solar powered turtle that can silently mow my lawn.
184.
▲
by
deckar01
3y ago
Not just high friction, but a formal verification like certification would be interesting.
185.
▲
by
deckar01
3y ago
I find that ingesting and chunking PDF textbooks automatically creates more of a fuzzy keyword index than a high level conceptual knowledge base. Manually curating the text into chunks and annotating high level context is an improvement, bu
186.
▲
by
deckar01
3y ago
Video: https://youtu.be/IwMgvRuqvBA
187.
▲
by
deckar01
3y ago
The account management / status tools were slow and flaky on the best days. I wouldn’t rule out a little extra traffic knocking them out. Correlation is not necessarily causation.
188.
▲
by
deckar01
3y ago
Evaluate is referring to measuring the accuracy of a model on a standard dataset for the purpose of comparing model performance. AKA benchmark. https://rentruewang.github.io/bocoel/research/
189.
▲
by
deckar01
3y ago
I ran into a GitHub issue template a long time ago that had a required section for alternate solutions. It’s amazing how often you find a better solution when you are forced to enumerate viable solutions and compare them.
190.
▲
by
deckar01
3y ago
My heuristic is how much noise is in the closest vectors. Even if the top k matches seem good, if the following noise has practically identical distance scores, it is going to fail a lot in practice. Ideally you could calculate some constan
191.
▲
by
deckar01
3y ago
It is two different language models. The embedding model tries to capture too many irrelevant aspects of the prompt that ends up putting it close to seemingly random documents. Inverting the question into the LLM’s blind guess and distillin
192.
▲
by
deckar01
3y ago
Instead of embedding the user prompt, I let the LLM invert it into keywords and search the embedding of that. It very much does feel like a magic bullet.
193.
▲
by
deckar01
3y ago
My LG OLED had intermittent contrast issues that persisted after toggling every permutation of picture bell and whistle. After finally stumbling across the correct search term, it turned out the offending feature was hidden in a maintenance
194.
▲
by
deckar01
3y ago
Large publishers could seek licensing deals similar to digital libraries. https://www.niso.org/niso-io/2014/12/reflections-library-lic...
195.
▲
by
deckar01
3y ago
Instagram is the same. They steal video from real products, funnel purchases through a random e-commerce page, and disappear by the time you need to request a charge back so that you can’t find evidence of the original offer.
196.
▲
by
deckar01
3y ago
The preview image shows intellisense covering up 80% of the code with irrelevant suggestions and a compiler error verbosely complaining about code in progress. That much distraction should have a very small visual indicator at the point of
197.
▲
by
deckar01
3y ago
Unfortunately, I think he still wants the skins. He is just tired of the formality of feeding and moisturizing his victims only to have them run away.
198.
▲
by
deckar01
3y ago
The assertion that DHS was funding the program (or using it as a tool) is an unsubstantiated leap. Stanford was gathering and analyzing intelligence with national security implications. There is an obvious public service in sharing this inf
199.
▲
by
deckar01
3y ago
It is misrepresenting the facts to the extent of being misinformation, which is quite ironic. It builds a grand narrative of conspiracy, but when the substantiating facts are finally sprinkled in they don’t quite match. > we just set up
200.
▲
by
deckar01
3y ago
Blu-ray is not raw, it is just HEVC with a higher bitrate. Most content encoded at a high profile level is not even going to hit the bitrate limit of a good streaming service + connection. Other than pausing during an action scene to take a
201.
▲
by
deckar01
3y ago
I would imagine the reliability would be higher. A resistive heating element could degrade and eventually develop a fault, but an inductive circuit decouples the electrical circuit from the heating element. A fault or an impurity in the hea
202.
▲
by
deckar01
3y ago
The Dolby Atmos/Vision standards also provide HDR, dynamic brightness, and surround sound. People spend thousands of dollars on home theater setups to meet those standards. Max has almost completely stopped producing cinematic content,
203.
▲
by
deckar01
3y ago
My AT&T internet package still “Includes Max” with no monetary value assigned to the “ad-free monthly” subscription, but new service with the same speed (without Max) is $5 cheaper monthly. Max’s subscription settings defer to AT&T,
204.
▲
by
deckar01
3y ago
There doesn’t seem to be a feature toggle for locking cameras in picture mode like there is for video. But now that I am fiddling with it, the icon that pops up to indicate auto-macro can be tapped to disable it… I find it bizarre that 1x m
205.
▲
by
deckar01
3y ago
The heuristic Apple uses for switching lenses is reminiscent of a street lamp that flickers on and off at dusk. How do they get away with marketing a device as professional that requires a 3rd party app to work?
206.
▲
by
deckar01
3y ago
The USPTO page on enablement is filled with examples of hardware and software patents that were rejected, because of hand-waving generalization. There needs to be some accountability for approving these bogus patents. https://www
207.
▲
by
deckar01
3y ago
Patenting a UI layout is even more rediculous. These companies are retracing the arbitrary decisions made to solve a problem, then widening the parameters to cover all possible solutions in a vague patent to use as a legal landmine for any
208.
▲
by
deckar01
3y ago
If you can’t implement a patent without spending millions of dollars on R&D filling in the blanks it isn’t an invention. This was an idea, and it shouldn’t have been granted a patent. *IMO NAL
209.
▲
by
deckar01
3y ago
It was interesting talking to my father, a former Christian minister, about AI. ChatGPT interactions had instilled some misconceptions and it was difficult to convince him that its responses were just cleverly weighted randomness. It produc
210.
▲
by
deckar01
3y ago
You don’t need time deltas for display. Just omit the prefix (left of the minute) that is equal to now.
More ›