aabhay

18 ms·

241.
by aabhay 2y ago
The article continuously mentions Google, but how does that affect Alphabet, its parent company? When I read about breaking up google, I had assumed that the division would simply force all alphabet companies to operate fully independently.
242.
by aabhay 2y ago
https://youtu.be/Wnb_n5YktO8?si=lCGrH36DvXATdyEs
243.
by aabhay 2y ago
Rust has some great and useful web frameworks that are a joy to use, once you understand what is going on. For example, in Axum, they use traits cleverly to allow you to use dependency injection the same way that fastapi uses it. But at lea
244.
by aabhay 2y ago
It seems to be making the argument that C++ 26 has new reflection features that outperform Rust code gen due to the unavailability of high quality proc macro features.
245.
by aabhay 2y ago
I don’t understand the argument. Rust proc macros have some level of introspection. I can retrieve the type information from the AST. The reason you can’t always know whether something has a trait is because that constraint isn’t implemente
246.
by aabhay 2y ago
The fact that most crates are compiled with serde as feature flag and the fact that you can raise a PR on most things that aren’t — are both great cases for Rust having an awesome ecosystem. Just as with async support, serialization support
247.
by aabhay 2y ago
We were in YC S22 and it definitely has become a status seeker magnet. However, entrepreneurship is itself a status seeker magnet. The point the author misses is that you need to think of YC not as an organization but as a segment of time f
248.
by aabhay 2y ago
Discounts
249.
by aabhay 2y ago
Indeed. We would have been sitting with our measly treasury bills, without a care in the world
250.
Poly (YC S22) is hiring Rust experts in SF to build "Arc browser for files"(ycombinator.com)
1 points by aabhay 2y ago
251.
by aabhay 2y ago
These are songs by Kendrick Lamar. I guess that these were used to try and add a youthful, lighthearted touch to the article? Didn't work on me though.
252.
by aabhay 2y ago
Power usage is indeed a better representation of GPU utilization during ML training. It has the advantage of combining many important indirect signals that aren’t visible, and avoids many downfalls of compute usage, which can go to 100% eve
253.
by aabhay 2y ago
If you’re looking for a Vue-flavored dev experience as opposed to a react one, I recommend this: https://github.com/antfu-collective/vitesse-webext Building a web extension is pretty straightforward in Vite-land. There
254.
by aabhay 2y ago
I think folks are misreading the article. Its not claiming that AI research isn’t valuable, its advocating for not pivoting into AI research. Research in any field requires a giant surplus of passion and talent. Go to where your passion and
255.
by aabhay 2y ago
I thought that this was going to use some interesting form of lambda calculus but instead it simply relies on the regex parser of find to compute things.
256.
by aabhay 2y ago
C’mon. OpenAI is a large company now with 1000+ employees. You’re really going to air this hot take? - if they release a model “they’re just releasing models without use cases” - if they release safety guardrails “they are just doing this t
257.
by aabhay 2y ago
Zoom out a bit. There’s a massive feeder ecosystem around llama. You’ll see many startups take this on and help drive down inference costs for everyone and create competitive pressure that will improve the state of the art.
258.
by aabhay 2y ago
The web based tools launch and move faster. There’s no lack of offline tools, if you’re the kind of person that files issue tickets in their spare time
259.
by aabhay 2y ago
Are self hosted runners down as well?
260.
by aabhay 2y ago
They’re likely paying for egress from the databases as well.
261.
by aabhay 2y ago
Very cool, I love that! I had a similar idea and it’s great to see that it’s worth exploring
262.
by aabhay 2y ago
In a future post, I would love to see a discussion of organizing your software abstractions, rather than just organizing files. For example, id love to know some ideas around keeping prior versions of structs around when you are partially u
263.
by aabhay 2y ago
Fair tradeoff if you ask me
264.
by aabhay 2y ago
If you need to run it continuously for a year
265.
by aabhay 2y ago
Yes but to place it in an unnecessary step between clicking “cancel” and “yes I’m sure, continue to cancel”?
266.
by aabhay 2y ago
Speaking from experience. More specifically, something like “whats the best brand of phone”. The LLM just summarizes common knowledge. But even a child will grasp some of the differences and have opinions drawn from experience. Note that th
267.
by aabhay 2y ago
I lived in Tokyo and literally didn’t know that the Yamanote line was so old, thanks HN! It runs so well, so much quieter than e.g. the BART. Insane that the loop was fully connected by the 1920s. If you do a bike/walk along the Yamano
268.
by aabhay 2y ago
To be honest I’m more surprised that this isn’t just how things are everywhere.
269.
by aabhay 2y ago
I think insert performance is a bad way to evaluate performance here, no? While B-Tree performance for time sorted keys is better on insert, what about during large transactions? In SQLite, my assumption was that the consensus was towards U
270.
by aabhay 2y ago
I have to be a soggy blanket person, but there are some pretty strong reasons why 3D generative AI is going to have a much shallower adoption cycle than 2D. (I founded a company that was likely the first generative AI company for 3D assets)

More ›