This is 100% true. Lidar improves accuracy by millimeters up close, inches at 10-50 feet away and feet beyond that. That accuracy is more than sufficient. Recognition and classification of objects is not improved at all (that part that matters). And, like parent post said, tesla classifies everything very very well, the real issue is that the planner acts completely crazy all the time and is scary.
There are two decisions when driving: go ahead around an obstacle or stop. Even as a human I do not need super high resolution to identify the objects asround, as long as I can identify if is in my path or not.
While our eyes can do that pretty reliable, we are organics and get tired - how many hours one can drive until this becomes almost imposible? I had a situation where I would hallucinate and and start believing something is in the street do I did a full stop - nothing happened, but was quite intense. Imagine the other way around - not stopping and hitting something.
A normal radar + some low level ASIC programming would do that without geting tired. My Audi from 2014 is quite good at that and I actually rely on this feature all the time.
Blackbird Studios | Sr Developers and Design-gineer | Remote in Americas or Western Europe | Contract to hire
Blackbird is a software dev studio specialized in a Human Centered approach to launching totally new, high-impact software products for a wide variety of public and private sector clients, from early stage startups to Fortune 50 companies. We don’t do marketing sites! We build complex, cool, interactive fullstack web and mobile applications that are fun to build, challenging, and that people use every day. We work directly with brands or through partnerships with some of the top design studios in the world (like IDEO).
Blackbird is very different from most other software studios. We operate in very collaborative and strategic relationships with our clients and partners which gives us a uniquely strategic seat at the table and the opportunity to regularly deep dive into a new interesting understanding of our client's business, then build, launch, and evolve new software products from the ground up.
Our services include CTO consulting, software development, UX/UI, training and capabilities building. We work with large and small companies/startups on projects ranging from 2-12 months with 2-10 people on each project to build full stack web, mobile, IoT, physical installations, etc. Our favorite stack is Typescript, React, React Native, Node (recently loving Remix!). We also use a wide range of other tech as needed since we face a pretty diverse set of challenges. Recent and upcoming examples: Flutter, Elixir/Phoenix, Rails, Swift, and Kotlin. We are a remote-first company but with very active collaboration on video and Slack. Occasional travel to San Francisco and other locations is part of the job for project kickoffs and other critical collaboration phases or meetings.
We are looking for people who are excited about building great software products, who are great teammates, and who are passionate about continuously refining best practices for creating successful digital products. Experience in dev team leadership, new product development, collaboration with business and design stakeholders, and consulting experience are not required but are all huge plusses.
Specifically, right now we are looking for:
- Sr full stack developers
- Lead level full stack devs with experience in team leadership roles
- A design-gineer.
For that last role, we're working on a cool internal design-tech project and need an awesome UX designer who also understands code and dev workflow.
Interested? We want to hear from you! Please send resume, Github, or LinkedIn to jobs@blackbird.us. I am a founder so ask me anything! If there are other aspects of what we do that interest you definitely still hit me up!
The design-gineer role: would you consider someone who started their career in design, works closely with designers, but has been predominately a full-stack developer for a decade?
I wouldn’t refer to myself as a designer today, but it’s in my roots, interests, and I still love to be involved in it. I’ve built a few basic visual simulators and editors, and it’s kind of my happy place; right between the heavy lifting of abstracting complex logic and the art and science making it look, perform, and feel great. If that’s anything like what the role entails, I’d love to chat.
When you consider the power that big oil and gas have worldwide, and all they've already done to sabotage adoption of clean energy, it just seems improbable to me that one day tech will arrive that provides unlimited clean energy without some kind of big ugly fight. Big. Like I can see these guys doing everything from run of the mill regulatory capture to kill it all the way up to supporting right wing (or communist) conspiracy theories or movements to destabilize democracies (all things that have been done in the past). I seriously wouldn't put anything past them. Maybe I'm being too paranoid but I have a hard time believing in any future that involves yanking away trillions of $$ in power from a small group of unscrupulous people.
this is the reason I've been saying that we will have fusion within a decade of when markets start to price in the decline of fossil fuels because of renewable & other factors. its not an impossible problem, it just needs more research funding/focus.
Part of the issue is that modern run-flat tires have stupidly low lifespans, like 20K miles. I have a Tesla and opted for non-run-flats with my latest change and I don't think my tires wear any faster than other cars I've had.
Elon is losing his mind. This is clear. Which is scary for a guy with so much power. My guess is in this case Adderal + sleep deprivation + living in an extremely distorted reality = mental breakdown.
Remix, zod, tRPC(if/when API endpoints are needed), Typescript, Prisma. This stack is gorgeous. Same language and type definitions front to back is a game changer.
The realtime aspect of what you are looking for is the hard part. The only options I know of are the remix stack above (you still have to do some extra plumbing), elixir/Phoenix (or other turbo links like options but most aren't fast enough). To do realtime updates natively you need to move to evented data front to back and that is a very big move away from models the way we normally think of them. Check out eventide or resolv.js.
> elixir/Phoenix (or other turbo links like options but most aren't fast enough
Phoenix specifically doesn't work like turbolinks. LiveVIew in Phoenix uses a websocket to transport data, morphdom to patch the dom, and a light process on the server to manage state. It's as fast as the internet connection for insertions into a list. In fact it is quite capable of parallel data loading just like remix, and the server can work share in parallel across as many cores as you have available on the server.
Probably a lot depending on your workload. Here's a 2015 article about getting to 2 million concurrent connections[1] with Phoenix channels. They used a bigger rack space box, but it was 7 years ago and Phoenix and Elixir have both improved a lot here. Here's another example of someone hitting a laptop with 840,000 connections[2] on an intel mac with the following specs:
create-t3-app is a great starting point for folks going this path. Has optional checkboxes for db/ORM/migrations (prisma.js), auth/sessions (next-auth.js), tailwind CSS, typed client/server communication (trpc).
I hadn't touched much webdev for a few years, the ecosystem / build tooling has gotten a LOT better. It's back to being fun and painless to spin up a new webapp.
A few hours ago, I tried to setup trpc using recommended boilerplate, I also tried fresh default T3 after that.
Both got error, which seems to be due to next/next auth dependency breaking. So it's not that painless I guess.. I found this situation quite often with nextjs example template
T3 author is very responsive though, so my issue resolved quickly
I can speak to a handful of things Remix needs to improve, but "makes things that were simple hard" is a very big generalization. Remix nets out for me big time. I've never been as productive with another framework. Also, remix does SSR but so do a lot of frameworks, is that what you mean when you say "__everything__ on the server again"??
I've heard this sentiment before (that AI != ML, always weirdly hostile) but I've also heard key figures in the AI world say strongly that ML of any kind is a subset of the broader AI umbrella. Are you sure it is strictly wrong to refer to this example as AI? Also why are you using such strong words? I'm genuinely curious why there is so much emotion when people, maybe, misuse these terms.
He's going for a middlebrow dismissal. This is like the "bitcoin isn't money!" or "twitter is not a serious medium for communication!" oldschoolism that some HN traditionalists defend in spite of real world application. At the end it doesn't uncover a salient point but gets stuck on semantics to shut down an entire idea.
You're not 100% wrong, but I'm not trying to shut down an entire idea of AI.
Let me put it this way: It's called AI, Artificial intelligence. Are trees inteligent?
Because they are a clump of cells that manage to achieve amazing results. growing to be some of the largest living things on the planet. Extracting nutrients and transporting them 20 or more meters above ground, where they are used to harvest energy from the sun, all the while producing offspring every year and fighting off predators.
I don't think many people would call them "natural intelligence".
The I admit that I maybe apply a narrow definition of "intelligence", but I think the core concept is one of "understanding".
And we are not even close to al ML algo actually understanding anything.
And this is the problem. It masks the inherent shortcommings of ML.
People are delivered the impression that applications that use ML actually do what they are expected to do, because like a person, you train them and then they understand their job and do it.
This is NOT HOW THAT WORKS. The ML algo does not understand that it is asked to identify oncoming traffic. It does not know that it is looking a cancer cells that will soon kill someone.
And even worse, we humans who make these things, are 100% unable to understand the models we create. We can feed them data, and compare the result. But that is it. There is no real way to understand how it works in detail.
They get used anyway. With predictable results. see Tesla Autopilot for a prominent example.
Also, I call them AI researchers for the same reason I call Nuclear Fusion researchers that, not because they are doing it, but because they are researching it.
I'm not the OP, but have been a data person long enough to hazard a guess.
If your job is an engineer or scientist, the term "AI" is basically a synonym for "unrealistic executive expectations." That's a super triggering and stressful situation to be in, especially early career.
I've actually had a CEO describe expected output of my ML team as "magic AI shit" - I bet you can imagine the team's reaction and tone. I'm reading the same strong emotions and frustration here.
The good news is you can always course-correct expectations with communication. I've come to love talking about AI with people who are only somewhat technical, because their wildest dreams are sometimes totally do-able with some duct tape and fine-tuning.
