The first thing Waymo tried building (way back when, circa 2010 or so) was highway based driver assist in the style of Autopilot. They did a ton of testing with it and didn’t like how quickly their testers stopped paying attention despite a ton of instruction not to. I’ve seen clips from these tests.
It’s also possible they shifted direction because the long term vision of robotaxis is much more lucrative.
The thing is, you can tune driver assist to whatever form you like; but then you have a true comparison of function with established competitors.
I just don't believe the approaches for high autonomy (especially at the time) actually could make a cost effective assist system.
And for whatever reason they decided to push the we didn't like the driver behaviour message, rather than actually talking about what was actually plausible to achieve in the driver assist space.
Unless you assume that self-driving software is perfect, no, it really isn't. That's the whole problem - the drivers would get complacent, so when there's an issue, they'd be caught by surprise and wouldn't be able to react.
Isn't the point to pay _no_ attention? The difference is when an accident occurs, was the person in the car at fault for not vigilantly watching everything.
You want the curve of total attention to be always above a baseline human in an unassisted car. The car can do some attention and the human can do some. But if the sum of the two falls below the threshold, you’re in trouble.
You can already do that by just closing your eyes and letting Jesus take the wheel. No, the point is doing so while maintaining safety.
It is materially less safe to operate a ADAS while distracted than driving manually. Humans are exceptionally good drivers on average, only encountering minor crashes on timeframes measured in years to decades. As such, if safety critical ADAS errors occur more frequently than every ~100,000 miles and you are attentive in less than 100% of all such occurrences, you are operating your vehicle multiple times more dangerously than the average driver (which is a number that includes drunks and distracted drivers).
That is why it is critical to deliberately downplay the capabilities, to avoid wishful over-reliance, and enforce strict driver awareness (through techniques such as driver monitoring) to avoid operating multi-ton killing machines in ways that are multiple times more dangerous to the occupants, other drivers, and pedestrians. Without that, people are prone to over-generalization of safety capabilities, extrapolating that a single success means robust, continued success thousands to tens of thousands of times in a row.
I don't think you'll ever be able to. Maybe not for decades for AR/VR. The AI stuff is cool; you can offload that, but mostly they'll just be a glorified camera mount and headphones for the actual 'glasses.'
Intels next gen iGPU has has similar benchmarks to a 3080. Compute is still scaling incredibly quickly. I think the optical problem will end up taking much longer than compute.
(phones are an obvious choice but I suspect you could jam significantly more power into the same form factor if you didn't need screen, speakers, mobile modem etc)
If anything, what I’ve heard from friends at or adjacent to Meta is that they’re paring back metaverse ambitions and capabilities on their future devices because of the success of the Ray-Bans glasses and the relatively middling sales of the Quest 3.
The Ray-Bans are also a weird anomaly since they’re leaning on the Ray-Bans designer pricing to justify a lot of the cost. If you’re already buying a pair of luxury sunglasses that cost close to $200, what’s another $100 to get the smart version?
And I keep my sunglasses in my car's glove compartment to use when I'm driving, with the only times I wear them being then and when I'm doing physical activity (running, sports, etc).
New, less physically risky use cases for sunglasses will not appear in my life just because I get a more expensive pair of them.
Don’t understand the appeal being a spy cam for Facebook, but to each their own. The subsidized quest is on borrowed time if it keeps losing crazy billions a quarter as it currently does.
What do you mean by “not great” grades? Failing, or not 4.0?
It’s honestly a little weird that you’re this into your child’s business after they’ve gone off to college, especially since this doesn’t even seem like a major crises (drug or alcohol problem, major psychiatric crisis) that I saw a ton of people go through in college.
Assuming they’re not failing, that they even got an internship after their first year shows a ton of initiative. I wouldn’t be too worried.
It does sound like you might be a helicopter parent though. In that case, did they get this internship largely of their own initiative or is it something you pushed them into?
> They are in a top ranked EECS program, think (Stanford, Berkeley, MIT, Cornell, CMU...) (getting a 3.0, and their parents still are upset about it)
All due respect, I too would black my mind out and escape and just want to play video games with my friends if I was from that kind of household/level of pressure/expectation.
I was equally gaming obsessed at around the same age. For me it was world of Warcraft. My parents were also concerned, but they showed interest in what I was doing and I explained it to them. When they realized I was leading of raiding Guild of over 50 people, and a good chunk of the time was actually me managing people, including resolving conflicts and all sorts of administrative tasks, they weren't nearly as concerned.
Depending on what game your son is playing (I've heard of Roblox but I'm not familiar with it), and how they're playing it, it very well may be that your son is developing some very valuable skills. It might not be managing people, but it might be creative problem solving, setting goals and working towards them, etc.
It sounds also like your son is a bit obsessive, and clearly very intelligent if he's able to achieve the grades he has while devoting so much time to a game.
I was this way too.
By the time I was 23 years old I tapped into some self motivation and desire for more financial freedom and became an entrepreneur. Through the rest of my 20 isn't into my early 30s that obsessive attention shifted from gaming and into my business. I didn't make millions of dollars but that type of focused effort led me to develop many valuable skills. I ended up selling my business for a small sum and now I make good money consulting on my own terms.
Frankly I think the most constructive thing you could do is just show an interest in what your son is doing in his game. In order for this to be constructive you have to reserve judgment. Don't hope for any specific answer just be curious and empathetic. Rather than reading about "gaming disorders" read about Roblox, maybe even try playing it a bit yourself, and try to really understand what it is your son is doing and why.
There are many paths to success. Almost no one follows exactly the path they set out for themselves in their youth, and most of those who do have mental health issues.
Assuming you want your son to be content and successful, support them in finding their own way, even if you don't fully understand it.
When I was 17 I had a career counsellor lecture me on there being no future in computers, and that instead I should become a carpenter or electrician, etc.
You don't know what the world will be like in 10 years or what skills your son will need to navigate it. You don't even know if college is what's best for him (even if you think you do).
So let go, let him make mistakes and figure it out for himself. He's clearly smart enough. He just needs you to be there for him and believe in him. Make sure he knows you love him unconditionally.
Dude your kid is fine. He’s going to a top tier college for his program and making a B average after his first year. Maybe he’s playing too much Roblox? Is he happy, healthy? Does he have a social life? If so, not a big deal in the grand scheme of things. I don’t even check the GPA of new grads I hire.
Edit: If it’s any consolation, I did know people who were truly addicted to gaming in college. They couldn’t manage their time and were on academic probation by the end of freshman year. Slept through multiple finals because they were up late playing league of legends. Never got any internships. Your kid doesn’t sound like the people I knew who really couldn’t manage their time. Also, Freshman year is generally academically the rockiest. A 3.0 is fine. The kids I knew who really couldn’t cut it were failing by the end of their first semester. And this is at a far less competitive program than where your kid is at.
It is also possible that the son may not really like CS as a subject and is using gaming to cope. I struggled to get above 3.0 in CS at a non-top school for CS but when I switched to a subject I liked (Economics) I was able to get almost straight As in undergrad, work as a RA in my last year, and get into a top 20 PhD program and succeed in that program.
Socialization is great, but too much socialization can be bad, too.
I didn't have enough alone time as a kid. I cannot exist alone.
I was always connected to the internet, so I would always be talking or interacting with someone. Always. I seemingly was not able to develop basic things that my brain should be able to do when alone, because I did not spend any time alone in order to develop them.
Now I can't do anything on my own unless I make it into a social thing. Some things by their very nature cannot be social things; I simply cannot do those things.
For example, I can't imagine things to myself. I can imagine things to describe to someone else, but I can't imagine things to myself. I can't entertain myself, either. Even watching videos or media, I have to also be talking to someone at the same time or I just get bored.
My only fulfillment comes from people. My happiness comes from making other people happy. My entertainment comes from entertaining other people. My imagination comes from depicting things to other people. My identity comes from showing other people.
But to myself, without anyone else, I do not exist.
Eh. There is a hole in me where "creativity" would normally be, because I would always just tell it to people instead of actually depicting it anywhere. As a result, it only really exists to be told to other people now.
Not sure if I'd call that "well-socialized" or "codependent". (although I guess codependence is usually on a specific other person rather than just people in general...)
I think it’s pretty clear that the OP meant that a downside of this is that you need a GPU to run it quickly locally, not that it being local first is a bad thing.
There's a lot of issues with it, but perhaps the biggest is that there aren't just troves of easily scrapable and digestible 3D models lying around on the internet to train on top of like we have with text, images, and video.
Almost all of the generative 3D models you see are actually generative image models that essentially (very crude simplification) perform something like photogrammetry to generate a 3D model - 'does this 3D object, rendered from 25 different views, match the text prompt as evaluated by this model trained on text-image pairs'?
This is a shitty way to generate 3D models, and it's why they almost all look kind of malformed.
If reinforcement learning were farther along, you could have it learn to reproduce scenes as 3D models. Each episode's task is to mimic an image, each step is a command mutating the scene (adding a polygon, or rotating the camera, etc.), and the reward signal is image similarity. You can even start by training it with synthetic data: generate small random scenes and make them increasingly sophisticated, then later switch over to trying to mimic images.
You wouldn't need any models to learn from. But my intuition is that RL is still quite weak, and that the model would flounder after learning to mimic background color and placing a few spheres.
We've had an awful experience with Next. There are undocumented, broken things all over the place. Migrating to the App Router was a huge mistake. 'use client' didn't work as expected, and we had to install workarounds and try out different hacks to get access to the most pedestrian features we could expect from any other web framework.
This is the broad consensus among everyone else I know using Next - they felt like the Next 13 release and the app router was a rug pull from a previously pleasant experience.
And whenever you complain about this or ask for support, some devrel guy from Vercel shows up in your replies, saying something to the effect of "Wow! That's crazy, we worked really hard on Next 13! This is my first time hearing about this!"
Edit: Since I know the devrel guys are in this thread: If you want people to keep using Next, something fundamental about how you guys write the framework needs to change. The instability and lack of documentation makes developing with your framework a massive pain in the ass - anyone I know who can is migrating off it as quickly as possible.
It's like the devs never tried building anything real with their own framework. Internationalisation support was dropped in the app router and the docs just tell you to DIY... but just to make the app support RTL languages you need to do some hacky things to update the <html dir="rtl"> in the root layout.
I’m sorry you haven’t had a good experience. And if you more specifics if anyone on the Next team replied to you in that way, please let me know.
I’ll be the first to say the launch of the App Router could have been smoother. Definitely not the first time I’ve heard the feedback. I’m optimistic Next.js 14 is a step in the right direction based on this.
What do you feel is undocumented? I’ll get that added.
Can you share a bit more about what you're trying to build? Is there a specific library you're wanting to have in Middleware versus putting it in a layout or page (which does support all of Node.js)?
Not trying to be offensive here and you're getting a lot of flak already, but these two issues are pretty self-explanatory...
It's a matter of configuration, it feels bad in the first one that a configuration is not possible if so desired and even worse having to jump through hoops in the second one because the edge runtime is enforced which one may not care about if all that's done for hosting is `next build` followed by `next start`
- more comprehensive documentation & tutorials around your own MDX solution would be great. It feels like you put that together, but don't intend anyone to really use it in production, expecting people to go for next-mdx-remote or mdx-bundler
- fix internationalization: I've heard multiple people complain that you now can't do mydomain.com/blog/first-post & have the French version at mydomain.com/fr/blog/first-post. Instead, you need to put everything under a locale, so the English version needs to be mydomain.com/en/blog/first-post
...just in general, it would be nice to see more documentation & tutorials for next 13 / 14 and the app router. People are also not very responsive on github, it feels like you are just moving forward, without waiting for anything to be stable.
I'm currently investigating the feasibility of using Next.js for a major refactor of our service and the one thing that stands out to me is how fractured the Next ecosystem is. More than 6 months after the App router was introduced the most important libraries are still in beta and don't have conclusive documentation for integrating with App dir.
For my projects localisation is a necessity. I tried both i18n and next-intl and found them both lacking in functionality, buggy and missing documentation. This should just be part of the framework or at least have a tighter integration.
The same story with next-auth.js, which confusingly still exists while promoting https://authjs.dev/. For the most basic implementation it probably works, but the app router documentation is spliced into the normal documentation which creates a whole lot of ambiguity.
There's been a lot of discussion surrounding caching for Next.js 13 as well. I personally find it confusing and the behaviour described in the documentation regarding revalidateTag/revalidatePath and client-side caching does not match my real world experience. I would love some more documentation regarding user-specific caching as well, for instance with personalised user dashboards.
It feels a bit ridiculous to release Next.js 14 today as we're still getting used to Next.js 13. And though there might not be any big/breaking changes it creates a feeling where Vercel is racing forward without keeping other library maintainers or its users in mind.
You can continue to use the Pages Router until there's broader ecosystem support for the larger React changes like Server Components. You don't need to use a library to handle localization or i18n.
i18n was part of the framework in the Pages Router, but it was limiting. We heard a lot of feedback that folks wanted better access to the raw primitives versus an opinionated i18n setup. So now you have full flexibility when using Middleware https://nextjs.org/docs/app/building-your-application/routin....
NextAuth.js just released a new beta version with full support for all App Router features, including Server Actions. It's what we're using in the official Next.js Learn course that teaches authentication https://nextjs.org/learn.
We mentioned this in the keynote today at Next.js Conf, but it wasnt in the Next.js 14 post, but next we're working on simplifying caching. We do now have extensive documentation on caching, but it kind of highlights that it's a bit much right now https://nextjs.org/docs/app/building-your-application/cachin....
Next.js 14 doesn't have new APIs to learn, so if you're learning Next.js 13 (which I believe you're referring to the App Router model), nothing changes. The major version is for semver, because there's a few small breaking changes like bumping the Node.js minimum version. We have some codemods https://nextjs.org/docs/app/building-your-application/upgrad....
I'm part of the team from inlang, we are developing solutions to make i18n both easier and more efficiently (e.g. with a web editor, vscode extension and CLI to automate your workflows). Our ecosystem is agnostic to every stack which means it can be integrated with every tech-stack out. We even have our own i18n library called paraglideJS.
Hard to say specifically, especially because I spent about a month trying to deal with this when we first migrated from Next 12 to 13 - but it essentially seemed that if you were using any web tech that wasn't built with SSR in mind, there were issues that were unaddressed in the documentation at the time.
Someone else in these comments mentioned:
- Fuck spending 3 hours working out why you’re not able to use relative image paths in MDX files and have to shove everything in /public.
- Fuck fighting five layers of configuration and bundlers and libraries and GitHub issues to try and load a WASM file without having the whole thing break.
These were both issues we had to deal with, and similarly lost many hours to trying to resolve.
At one point you guys seemed to have shipped a release that, for some large fraction of users, caused the devserver to start OOMing rapidly. When this started happening to me, I spent about a day trying to debug it, since I assumed that such an a huge and imminent issue would arise from my mistake, not the web framework's. I eventually texted a friend of mine about the problem, only for him to tell me that they had been having the same issue, and linked to this thread.
Working with a web framework that's so unstable that fundamental features like the devserver break periodically isn't fun. I stop trusting the web framework, and when a bug arises I find myself having to check both my own code and Next's. Now this applies to some degree with any open source framework, of course - there's bugs in everything. But I have so little trust for Next at this point. We've collectively lost weeks of eng hours to just dealing with quirks of Next.
Right now we're really affected by the fact that the recommended way to cache database connection objects is broken in our repo (and some other users' repos, evidently).
This is another example of something where I naively assumed that the error was on my end, and spent days trying to debug our database configuration, hosting provider, etc, before realizing that at some point something broke and the recommended way of persisting these connections on the global object doesn't work anymore. What's the point of using a framework if I have to question whether it's working correctly every time an error arises?
I gotta say, your responses in this thread more or less typify the interactions with Vercel devrel that I complained about in my parent comment. Every time I, or someone I know, complains about Next or try to raise an issue on social media, you or someone else at Vercel shows up asking for more info. If we give you any info we get a polite "Ah thanks for letting us know, yeah that sucks!", and nothing is done, as far as we can see, to address the issues we brought up or course correct Next's development.
I plan to look into the issues that you mentioned in your previous comment. I'm sorry that I can't get to it today, but I will follow up on those issues.
Correct, I'm not sure what problem you're having but here's an upbeat response nicely stating it's not Vercel's fault, and/or explaining why it's not a real problem.
Honestly, I've had a great experience because I've read a lot of experiences like yours and stuck with pages router & Next 12, which works nicely for everything I've needed it to do.
There might be a point where app router is stable & smooth but it's pretty clearly not right now, so havn't really seen the need to upgrade. I think there was a pretty decent comms issue with the stability of it from both the Next and React teams, but I have a hard time faulting an otherwise fairly stable and useful framework for adding features when they're not breaking the existing stable path.
Hooks was a bit of a bumpy transition as well, but I do think that I prefer the code written with them to the code before them. I think it's OK to wait a year or two to let the rough edges get filed down when these types of frameworks release big new feature sets.
Edit: I'll note that we don't use next/image or API routes either, both of which I've seen some churn / pain with. Possible I just hit on the framework when it was in a pretty happy place and most of the new features or suggested defaults have had pain points that I havn't experienced.
Remix is like the Next.js Pages Router + a helper similar to Server Actions. It does not support Server Components (what the App Router uses) and they're exploring using RSC in the future. So definitely different.
For example, `getServerSideProps` in Next.js is the same as a `loader` in Remix.
Astro probably, if they want to keep using React components.
Otherwise they might also be looking at Nuxt, Sveltekit or Solidstart, but if they're switching top-level frameworks they may as well use Astro because they can mix and match React/Vue/Svelte/Solid
Seems like Cruise doesn't actually have this. If so, why aren't we hearing a report from the remote driver? Why did the remote driver let the car drag a woman? Why do remote drivers let Cruise cars screw up over and over again, block emergency vehicles, etc...
If I'm right, then Cruise will have to add remote drivers and then we can call them remotely operated cars instead of driver-less cars.
They have this, by their own admission and from my personal experiences in the car. A detail you might be mistaken on - they’re not really fully driving the car remotely. They’re basically sitting in a call center answering questions for the car or plotting a course, or giving it the all clear to proceed if things are uncertain. And generally this only happens if the situation is one where the car can ‘safely’ stop and wait for a response. In situations like the one discussed here, the car doesn’t have time to wait for a response from a human operator, so it just does what it thinks is best locally.
It’s also possible they shifted direction because the long term vision of robotaxis is much more lucrative.