jowday

10 ms·

61.
by jowday 4y ago
They’ve also got real ones with no safety driver in San Francisco right now.
62.
by jowday 4y ago
It’s not. See https://en.m.wikipedia.org/wiki/Showrunner
63.
by jowday 4y ago
He’s putting those in quotes because they’re generally not formal positions that are explicitly credited, but are informal phrases referring to whoever has final creative control over a series.
64.
by jowday 4y ago
Sure - the MCU uses a lot of TV directors for example - since it functions more like a TV show than a auteur-driven series of films.
65.
by jowday 4y ago
> Looking at the imdb page this seems like a corporate driven project rather than a creator driven project with lots of different directors that haven’t don’t anything remarkable and seem to have just done other Amazon shows. With the ex
66.
by jowday 4y ago
The “ethical concerns” thing is just a progressive-sounding excuse for why they’re not going to give their models away for free. I guarantee you those models are going to be integrated into various Google products in some form or another.
67.
by jowday 4y ago
I mean I’d count adding lane markers as building infrastructure - and the same caveat of people disobeying them applies. Not to mention you wouldn’t be able to add dedicated lanes outside of main thoroughfares and highways.
68.
by jowday 4y ago
Requiring dedicated infrastructure eliminates the point of self driving cars - you might as well just build public transit infrastructure. On top of that, this won’t eliminate any of the real edge cases - humans and other cars will still vi
69.
by jowday 4y ago
This is something that robotaxi companies have already spent a ton of time figuring out - there are a ton of different metrics the cars have to hit before they remove the driver (as companies have done in Phoenix, San Francisco, Miami, and
70.
by jowday 4y ago
Much like OpenAIs marketing speak about withholding their models for safety, this is just a progressive-sounding cover story for them not wanting to essentially give away a model they spent thousands of man hours and tens of millions of dol
71.
by jowday 5y ago
The nerf already does this - the problem is when part of the object doesn’t appear in any of the frames.
72.
by jowday 5y ago
You’ll find this is true of all NeRFs if you spend time playing around with them. If a NeRF is trying to render part of an object that wasn’t observered in the input images, it’s going to look strange, since it’s ultimately just guessing at
73.
by jowday 5y ago
This sort of stuff (generating 3D assets from photographs of real objects) has been common for quite a while via photogrammetry. NeRFs are interesting because (in some cases) they can create renders that look higher quality with fewer photo
74.
by jowday 5y ago
It’s strange to me that the article focuses on Requiem for a Dream when Black Swan is much more directly inspired by Perfect Blue - if you wanted to you could even argue that it’s a remake of Perfect Blue. The plots are that similar. The ar
75.
by jowday 5y ago
Gravity’s Rainbow
76.
by jowday 5y ago
> There would never be a mix of self driving car and human driving car. There already are in Phoenix and Arizona. >There will be self driving cars on special self driving lanes. This doesn’t solve any hard problems for AVs. Rules arou
77.
by jowday 5y ago
A level 5 car that only works in major cities is, by definition, not a level 5 car. A level 5 car can drive everywhere in all conditions.
78.
by jowday 5y ago
A level 4 car that’s dependent on HD maps, remote support, and constrained ODDs is not a level 5 car. A very safe level 4 car wouldn’t fulfill Carmack’s side of this bet.
79.
by jowday 5y ago
True level 5 is essentially AGI - or at least, if we’ve achieved level 5 AVs in a hypothetical future, I’d say it’s a safe bet we’ve also achieved AGI. On that alone I’d bet against Carmack. Level 4 ridehailing in all major cities would be
80.
by jowday 5y ago
Confusingly enough, Cruise has nothing to do with Super-Cruise. Super-Cruise is handled by an in-house team at GM.
81.
by jowday 5y ago
I don’t think image+text pair models (E.G. CLIP) would be very useful for AV tasks. They’re not very good at fine-grained classification or counting the number of instances in a scene. Not even getting into latency or model size.
82.
by jowday 5y ago
I mean, I described the reason. The other thing is that perception the way you’re thinking about it isn’t a huge problem for AVs. They’re generally able to recognize all of the objects and elements they need to from their onboard sensors. T
83.
by jowday 5y ago
The sort of abstract reasoning I’m talking about is beyond the capabilities of any ML model that will run onboard a car - the “abstract reasoning” problem in AVs right now is solved primarily through HD mapping and remote assistance. LIDAR
84.
by jowday 5y ago
And I’m saying it doesn’t make sense to do that with Cruise’s stack.
85.
by jowday 5y ago
There are tons of examples of objects flickering in and out of existence, not being detected, warping in weird ways. This is one of the most recent tweets from a prominent Tesla reverse engineer with plenty of examples. https://t
86.
by jowday 5y ago
It’s very likely those LIDAR vehicles are used to gather ground truth to train depth models or calibrate sensors.
87.
by jowday 5y ago
The ADAS system in the Escalade has nothing to do with Cruise (confusingly). It’s developed by an in-house ADAS team at GM. Unlikely Kyle wants to become CEO of GM. Until recently the message seemed to be that Cruise was going to operate as
88.
by jowday 5y ago
Tesla isn’t focused on robotaxis - they’re focused on trying to sell a premium ADAS feature their CEO over promised half a decade ago. Whcih is necessarily designed very differently from a robotaxi. It doesn’t make sense to try to force a s
89.
by jowday 5y ago
Define “fully autonomous” - we have driverless cars in Phoenix and San Francisco right now - granted, with some pretty narrow constraints. Not as narrow as “needs custom hardware installed along every road to work” Another thing to underst
90.
by jowday 5y ago
> Aren't AI models already better at image recognition than humans? On some benchmarks, AI models are better at very well defined tasks like image classification (“label this image from a set of 8 labels you’ve seen before”) or obje

More ›