_ea1k

16 ms·

271.
by _ea1k 2y ago
> Why is this hard to accept? Because a lot of people on the other side of the aisle from the current executive said it is bad. And then they used ad hominem attacks and random slanders to try to shout down anyone who says otherwise. It&
272.
by _ea1k 2y ago
The good thing is that many of these are lifepo4, which largely mitigates the risk.
273.
by _ea1k 2y ago
Indeed, the new title is far better. Thanks!
274.
by _ea1k 2y ago
That's what I thought too, but apparently the title is pure, absolute, rage-inducing clickbait. The actual conclusion is that they make classes of errors that traditional OCR programs either don't make, or make in different ways.
275.
by _ea1k 2y ago
You'd think they would include this kind of detail in the training set, but it gets overlooked.
276.
by _ea1k 2y ago
Yeah, and they'll likely make this Appley only to create social pressure for even more uptake. Yuck
277.
by _ea1k 2y ago
I feel like a lot of people in this thread have never done continued training on an LLM and it shows. Seriously, a set of weights that already works really well is basically the ideal basis for a _lot_ of ML tasks.
278.
by _ea1k 2y ago
I've been poking at Uber forums and other places, and I'm noticing that a _lot_ of people are underestimating this. Waymo is growing by >100%/year at the moment and this is a significant part of that growth. It'll mak
279.
by _ea1k 2y ago
That's probably why he suggested a roth. $7k might be too much, but you can withdraw your own contributions from a roth without penalty. This makes it a great way to save for people early in their career, especially if they are in lowe
280.
by _ea1k 2y ago
That's probably why he suggested a roth. $7k might be too much, but you can withdraw your own contributions from a roth without penalty. This makes it a great way to save for people early in their career, especially if they are in lowe
281.
by _ea1k 2y ago
At the other end of the spectrum, the belief that the moon landing was faked seems to be steadily increasing in popularity. I don't get it at all.
282.
by _ea1k 2y ago
No, he means wait behind the line.
283.
by _ea1k 2y ago
I had the same thought. I've been in places where the first one would supposedly get you a ticket, as you are intended to stop if safely able to do so. It looked like that might have been the case there, though it isn't clear at a
284.
by _ea1k 2y ago
I worked at a government site with a government machine at one time. I had an issue, so I took it to the IT desk. They were able to get that sorted, but then said I had another issue. "Your CPU is running at 100% all the time, because
285.
by _ea1k 2y ago
I think you need to have a pretty good understanding of the underlying workings of at least a couple of modern models. You should also follow along closely enough to see if major changes happen in the way they are built. But watching every
286.
by _ea1k 2y ago
Yeah, there really needed to be some voting or other feedback system to encourage consistency over time. This is a mildly interesting experiment that largely just looks like a failure at the moment.
287.
by _ea1k 2y ago
Yeah, it'd likely need 150kwh or so. That'd probably make it a $70k product, and I doubt the market for $70k minivans is huge. Then again, that is about the price of the id.buzz and some find it appealing with its much smaller pac
288.
by _ea1k 2y ago
> Let that sink in: people were so angry at their property, health, and safety put at risk as unconsenting "beta testers" that they set fire to one of the cars. Maybe the sort of person that would take that action isn't a
289.
by _ea1k 2y ago
I really like the ~3B param version of phi-3. It wasn't very powerful and overused memory, but was surprisingly strong for such a small model. I'm not sure how I can be impressed by a 14B Phi-4. That isn't really small any mo
290.
by _ea1k 2y ago
It isn't really hypothetical. Lots of good models run well on a modern Macbook Pro.
291.
by _ea1k 2y ago
Yeah, people act like some Tesla problems are bad, but I've certainly used systems that are worse. Just in a few miles, I've had a Nissan accelerate towards stopped traffic aggressively enough to trigger its own collision warning.
292.
by _ea1k 2y ago
TBH, I'm finding that people underestimate the usefulness of CPU in both inference and fine tuning. PEFT with access to 64GB+ RAM and lots of cores can sometimes be cost effective.
293.
by _ea1k 2y ago
My dealer experiences have included hours to swap out a dome light under warranty, 1+ hours to do a software update, and a dealer trying to claim that the manufacturers automatic engine warranty extension didn't apply to my car. Also,
294.
by _ea1k 2y ago
I'm guessing that 7 comes from splitting J1772 from CCS1. It sort of makes sense, I guess. There are even a few EVs with J1772 but not CCS1 (eg, the early Chevrolet Bolt without the DCFC option). I agree with you, though. Splitting it
295.
by _ea1k 2y ago
People say that, but just like with other cars, it is very vehicle, timing, and trim dependent. The Rivians seem to be doing ok. Even a lot of Teslas are doing ok, not great. Those Model Ys that people bought for ~70k a couple of years ago?
296.
by _ea1k 2y ago
A2Z has announced an extension cord, but tbh it sounds like a bad idea: https://a2zevshop.com/products/dc-nacs-extension-cord-6-9-ft... That's a lot of power through an air cooled cable.
297.
by _ea1k 2y ago
All CCS cars support 400V charging, though. Just some of them do it much more slowly compared to an 800V charger. This is why Tesla is able to work with CCS, despite superchargers not supporting 800V.
298.
by _ea1k 2y ago
That's true, but there's also another weird counteracting effect. Lets say you have 4 chargers offering a full 350 kw each. If they get used at full power for ~15 minutes in the month, the demand charge for the month might be ~$10
299.
by _ea1k 2y ago
Yes, just use "detect ball" as the prompt. It will give you y1,x1,y2,x2 coordinates on a scale of 0-1024. It is really good at this. Unfortunately, without fine tunes you can't have it just detect everything and return all de
300.
by _ea1k 2y ago
Government overreach is the big thing that I'm worried about. All of the overreactions to adas is really heavily driving it. At some point, they will have to realize that the same attention systems used for adas could be used all of th

More ›