Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
danjl
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
25 ms
·
331.
▲
by
danjl
2y ago
Fortunately, most text is proportionally spaced, rather than monospaced, so you cannot really determine the number of characters, only the length in pixels/points of the redacted phrase.
332.
▲
by
danjl
2y ago
Exactly, the stupid bit is using the original text to drive the pixelization. Another approach would be to just generate random gray pixel values over the redacted text. Simple. Simpler even than the weird assumption that you would use the
333.
▲
by
danjl
2y ago
Certainly still worth learning CNNs. Still unclear if ViT is better. And there's certainly enough for a full course on CNNs and a separate course on vision transformers.
334.
▲
by
danjl
2y ago
Even better with a dog
335.
▲
by
danjl
2y ago
Exactly.
336.
▲
by
danjl
2y ago
You seem to be inflating the emotional importance of my comment. Google did an enormous amount of the research prior to scaling. I merely pointing out that if there's credit to be given out, a bunch of it goes to Google.
337.
▲
by
danjl
2y ago
Open Source != Free Software There are thousands of companies who release open source as part of a paid product.
338.
▲
by
danjl
2y ago
While you're at it, stop looking for VCs and investment to pay your salary. Instead, build under revenue. Bootstrap. So many of the problems in startups happen because people want different things. If you don't have a co-founder,
339.
▲
by
danjl
2y ago
In that vein, I'd say that most of the LLM research was done at Google. OpenAI productized faster.
340.
▲
by
danjl
2y ago
Check out https://1lilac.com . Still in development. Pricing is temporary. Keen to hear feedback. This likely falls into the "vastly over-engineered" category. Also, it is only online at the moment. Disclaimer: I am the
341.
▲
by
danjl
2y ago
I bet traditional image processing would love to be implemented in ThunderKitten.
342.
▲
by
danjl
2y ago
I remember the hype around Google Wave, but I never personally had the opportunity to try it out. I recall seeing some screenshots that looked a bit like Google groups but with more icons. I did use Google Circles and Google+. Sounds like
343.
▲
by
danjl
2y ago
Heh, perhaps you could use the text comments to constantly update a visual description generated by AI? That way, nobody needs to make a drawing. Conversely, nobody has direct control over the visuals. I envision a UI where people type thei
344.
▲
by
danjl
2y ago
Visuals are the big missing piece. Any text-based discussion will build different visions in each reader's head. Often, everyone agrees on the text description, and then, when a designer draws a picture, everyone disagrees because they
345.
▲
by
danjl
2y ago
Do kids these days actually make Daisy chains anymore? You know, with real Daisy flowers.
346.
▲
by
danjl
3y ago
The vertical layout is simple to implement. Horizontal is much harder, and more useful for media like images and videos, see Flickr. None need CSS. And, if you are going to do it in CSS, you better handle both, not just the easy less useful
347.
▲
by
danjl
3y ago
The problem is not writing code. The problem is writing the right code and supporting it as the product evolves. AI tools result in programmers that know less about their code and therefore less able to support and less able to contribute t
348.
▲
by
danjl
3y ago
Managers are just the symptom. Growth is the disease. There is no good solution to managers except being small enough to not need them at all. The idea that companies need to grow to survive is a myth. Staying small avoids so many problems.
349.
▲
by
danjl
3y ago
There are many more desktop tools for debugging GPU code, largely because of game development. Xcode as an IDE is pretty out of date. Visual studio has much better debugging tools for GPUs. See nvidia's tools. As a long time GPU progra
350.
▲
by
danjl
3y ago
This reads like promotional literature for using Cloud services. Why do all this crap when you can just use AWS or GCP? Heck, as a bonus, you can blame them when it breaks instead of getting fired!
351.
▲
by
danjl
3y ago
Speaking for those of us who were born and raised here, please don't come. We have enough. Nobody who has been here for any reasonable period actually left.
352.
▲
by
danjl
3y ago
I like the "unsafe function" idea. More generally, it would be great if you could somehow show an "unsafe chance" metric, which would show lines that can cause multithreading issues. That is, rather than relying on the u
353.
▲
by
danjl
3y ago
Love it. I want to color the lines based on their chance for causing errors. It would have to be a heuristic based on various metrics, like the number of historical changes to a line, code complexity, and test/coverage hits.
354.
▲
by
danjl
3y ago
No, search -> ads. It isn't like their ads come from Google Meet or GMail. Their ads come from search. Their search is still best with no challengers of note.
355.
▲
by
danjl
3y ago
Terrible writing. Creates drama where it doesn't exist. The author makes statements about "problems at Google" and provides anecdotal evidence. Desperately tries to create a "knight on a white horse" storyline for n
356.
▲
by
danjl
3y ago
Aha! A free domain name generator! Not sure why this post was mis-titled?
357.
▲
by
danjl
3y ago
This is the curse of the enterprise. Small companies can avoid these issues entirely, if they want. And, if they don't and you find a startup with lots of titles like CxO and Head of LaDiDa, run for the hills since they have inexperien
358.
▲
by
danjl
3y ago
I just don't believe this story. It feels like something blown out of proportion and then seeded to the press.
359.
▲
by
danjl
3y ago
I would guess that Google's vertexAI managed solution uses TPUs. Also Google uses them internally to train and infer for all their research products.
360.
▲
by
danjl
3y ago
Isn't this basically, "how to structure a large project". The bits that are C-specific are small. This could easily be changed to JavaScript, Rust, Java or any language with only small modifications.
More ›