Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
yunwal
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
42 ms
·
751.
▲
by
yunwal
4y ago
True, but previous chips have been manufactured in China, and they’re also developing and manufacturing their successor to A100s (H100s) in China. https://www.cnbc.com/2022/09/01/nvidia-says-us-government-al..
752.
▲
by
yunwal
4y ago
> Also making a queue bigger is a bad reflexive response to queues being full. Depends on the function of the queue. If it's a dead-letter queue, there's literally no downside besides a likely-trivial amount of cost. If the pro
753.
▲
by
yunwal
4y ago
> It's quite misleading for FB to claim that they are being so open, but then hoard the weights and only release it to a few academics. Facebook almost certainly knew leaks would happen. My guess is keeping the model "contained
754.
▲
by
yunwal
4y ago
Long term I also imagine it's much cheaper to run these large model trainings on renewables. It's a very centralized process that doesn't necessarily need 100% availability. The manufacturing process, however, is totally dece
755.
▲
by
yunwal
4y ago
Being a good manager is in my experience just as hard if not harder. But, it's much easier to be a bad manager and keep your job.
756.
▲
by
yunwal
4y ago
> People with real experience will have encountered this problem and know that you have to spend time investigating to find the root cause of the backpressure, otherwise you could add capacity to random things without reducing the queue
757.
▲
by
yunwal
4y ago
aws-vault[0] does this, but only for aws creds 0 - https://github.com/99designs/aws-vault
758.
▲
by
yunwal
4y ago
I don’t know how to homebrew. If I ask ChatGPT to help me get started homebrewing, it lists helpful steps to start homebrewing. I can ask it to expand on any of those steps until the breakdown is actionable. Checking some of the facts it gi
759.
▲
by
yunwal
4y ago
Nothing wrong with it but it’s pretty devoid of nutrition compared to other greens.
760.
▲
by
yunwal
4y ago
I mean the obvious counterexample is someone who eats 1500 calories of real food and 2500 calories of junk food a day. If you scale that down by half you start getting all sorts of nutrient deficiencies well before hitting your target weigh
761.
▲
by
yunwal
4y ago
> It absolutely can It can’t. There are many people who, if they simply scaled down their diet without otherwise modifying their nutrient intake/lifestyle, start having health problems well before they lose weight. Hot flashes, fain
762.
▲
by
yunwal
4y ago
> So...what's the moat? Why can't anyone build a clone sight and come in with slightly cheaper shipping and undercut his business? The dumbest dismissals of business ideas always start with this phrase. Moats aren't like s
763.
▲
by
yunwal
4y ago
I've been thinking about this recently and I think the answer (for me) is to get a non-tech job/full-time-hobby for a bit. It makes sense that all the ideas in the "corporate office/tech infra" space are all used up
764.
▲
by
yunwal
4y ago
These remind me of the posters you’d buy in weed dispensaries that came with a free “gift” in places like DC where possession was legalized but sale was not.
765.
▲
by
yunwal
4y ago
Not the only one, but idk why people would follow Musk’s lead as opposed to, say, Tim Cook’s, when apple’s valuation has barely taken a dip.
766.
▲
by
yunwal
4y ago
> Nonetheless, I think that being cynical about anything Amazon is up to isn't unwarranted. 100%. I still don’t think having the 21st century standard-bearer for anticompetitive practices move into healthcare is a good thing by any
767.
▲
by
yunwal
4y ago
I mean, fidelity has no reason to cast false aspersions at musk especially since they own equity in several of his companies https://www.axios.com/2023/01/29/fidelity-cuts-twitter-valua...
768.
▲
by
yunwal
4y ago
Financial institutions that own significant equity in twitter still publish their valuations https://www.axios.com/2023/01/29/fidelity-cuts-twitter-valua...
769.
▲
by
yunwal
4y ago
“High crime” urban areas are still generally safer for kids than car-centric suburban areas because cars aren’t going as fast. Kids are far more likely to die getting hit by a car than from violent crime.
770.
▲
by
yunwal
4y ago
> If you scrap together an AWS migration tool that can pull databases, instances, and lambdas Ah yes, I’ll just throw a quick script together. Never mind that there are entire consulting firms dedicated to doing exactly this and each of
771.
▲
by
yunwal
4y ago
I mean, twitter’s worth about 40% of what it was when Musk bought it. Hardly a success story.
772.
▲
by
yunwal
4y ago
> Not everyone who who handles medical data is bound by it This is true, but not true of anyone working at One Medical, since health care providers are bound by HIPAA. If you tell your friend you have cancer, sure, they can legally share
773.
▲
by
yunwal
4y ago
Medical data, despite seeming like a juicy target, isn't really all that valuable atm. It requires a lot of infrastructure to move it around, and you can't do all that much with it without breaking the law. HIPAA/PHI fines ca
774.
▲
by
yunwal
4y ago
Using pre-trained AI models only requires surface-level knowledge about AI. You need to know its limitations, but you don't need to know tensorflow or how transformer networks work to be able to use AI for consumer applications.
775.
▲
by
yunwal
4y ago
This is the only advice that worked for me. Before you settle down to watch TV for the night (or read a book, or whatever it is you do to unwind), prep yourself for everything you want to get accomplished the next day. Pack your gym clothes
776.
▲
by
yunwal
4y ago
> I find it funny that when you drill down, the only supposed quality decrease is in the "freshness" of produce and meat I mean, these "subjective" things are basically the only thing Whole Foods has going for it. Eve
777.
▲
by
yunwal
4y ago
Amazon's likely buying One Medical for the same reason as most of it's acquisitions: to have an in-house team to dogfood AWS solutions. They'll care about customer experience enough to maintain a reasonable customer base, but
778.
▲
by
yunwal
4y ago
I’d like to be able to search long videos (or podcasts) for a keyword (or even better, semantic search), and be directed to all time stamps in that video that match. Things like this sort of exist, but nothing that’s really usable.
779.
▲
by
yunwal
4y ago
But people do seem to think that just because ChatGPT doesn't do movie listings well, that means it's useless, when it is perfectly capable of doing many other things well.
780.
▲
by
yunwal
4y ago
> It's possible the prompts are generic enough that they work well against all AIs The skill here is to be able to systematically understand language of different AIs and how to instruct them. The various "tips and tricks"
More ›