This is a bit snarky — but will Intel actually keep this product line alive for more than a few years? Having been bitten by building products around some of their non-x86 offerings where they killed good IP off and then failed to support it… I’m skeptical.
I truly do hope it is successful so we can have some alternative accelerators.
The real question is, how long does it actually have to hang around really? With the way this market is going, it probably only has to be supported in earnest for a few years by which point it'll be so far obsolete that everyone who matters will have moved on.
We're talking about the architecture, not the hardware model. What people want is to have a new, faster version in a few years that will run the same code written for this one.
Also, hardware has a lifecycle. At some point the old hardware isn't worth running in a large scale operation because it consumes more in electricity to run 24/7 than it would cost to replace with newer hardware. But then it falls into the hands of people who aren't going to run it 24/7, like hobbyists and students, which as a manufacturer you still want to support because that's how you get people to invest their time in your stuff instead of a competitor's.
What’s Next: Intel Gaudi 3 accelerators' momentum will be foundational for Falcon Shores, Intel’s next-generation graphics processing unit (GPU) for AI and high-performance computing (HPC). Falcon Shores will integrate the Intel Gaudi and Intel® Xe intellectual property (IP) with a single GPU programming interface built on the Intel® oneAPI specification.
I can't tell if your comment is sarcastic or genuine :). It goes to show how out of touch I am on AI hw and sw matters.
Yesterday I thought about installing and trying to use https://news.ycombinator.com/item?id=39372159 (Reor is an open-source AI note-taking app that runs models locally.) and feed it my markdown folder but I stop midway, asking myself "don't I need some kind of powerful GPU for that ?". And now I am thinking "wait, should I wait for `standard` pluggable AI computing hardware device ? Is that Intel Gaudi 3 something like that ?".
I think it's a valid question. Intel has a habit of whispering away anything that doesn't immediately ship millions of units or that they're contractually obligated to support.
I'm not very involved in the broader topic, but isn't the shortage of hardware for AI-related workloads intense enough so as to grant them the benefit of the doubt?
Itanium only failed because AMD was allowed to come up with AMD64, Intel would have managed to push Itanium no matter what, if there were no alternatives to a 64bit compatible x86 CPU.
Itanium wasn't x86 compatible, it used the EPIC VLIW instruction set. It relied heavily on compiler optimization that never really materialized. I think it was called speculative precompilation or something like that. The Itanium suffered in two ways that had interplay with one another. The first is that it was very latency sensitive and non-deterministic fetches stalled it. The second was there often weren't enough parallel instructions to execute simultaneously. In both cases the processor spent a lot of time executing NOPs.
Modern CPUs have moved towards becoming simpler and more flexible in their execution with specialized hardware (GPUs, etc) for the more parallel and repetitive tasks that Itanium excelled at.
I doubt that Itanium would have ever perked down to consumer level devices. It was ill suited for that workload because it was designed for highly parallel work loads. It was still struggling with server workloads at the time it was discontinued.
At Itanium's launch, an x86 Windows Server could use Physical Address Extension to support 128GBs of RAM. In an alt timeline where x86-64 never happened, we'd have likely seen PAE perk down to consumer level operating systems to support greater than 4GB of RAM. It was supported on all popular consumer x86 CPUs from Intel and AMD at the time.
The primary reasons we have the technologies we have today was wide availability and wide support. Itanium never achieved either. In a timeline without x86-64 there might have been room for IBM Power to compete with Xeon/Opteron/Itanium. The console wars would have still developed the underlying technologies used by Nvidia for it's ML products, and Intel would likely be devoting resource into making Itanium an ML powerhouse.
We'd be stuck with x86, ARM or Power as a desktop option.
I haven’t read the article but my first question would be “what problem is this accelerator solving?” and if the answer is simply “you can AI without Nvidia”, that’s not good enough, because that’s the pot calling the kettle black. None of these companies is “altruistic” but between the three of them I expect AMD to be the nicest to its customers. Nvidia will squeeze the most money out of theirs, and Intel will leave theirs out to dry when corporate leadership decides it’s a failure.
Darkhive.com | United States | REMOTE | Full-time | $150-220k + early shares, employee #10-15
There are two kinds of drone companies on the market. DJI and everyone else you haven't heard of. If you want to change that, particularly in the US, check us out. We are creating a palm-sized, autonomous drones to help de-escalate dangerous situations.
If ROS2, sUAS, GTSAM, SolidWorks, aerospace, hard realtime systems, and drones are things you're familiar with or interested in, let's chat. Shipped a consumer product in volume? Let's chat.
Maybe you're working an another UAS/Drone company that's larger and want to flex your brain to have more direct impact and ownership of your creations. Apply below and mention HackerNews.
Let's make drones more affordable and accessible to public safety, humanitarian, and defense applications.
- Position can be 100% remote (US Only). Relocation to San Antonio, TX would be awesome.
- We do in person meetings when needed. We are just 14 people.
- We're a startup, but we have funding and contracts. We're all mid-career so we want a sustainable, mature, diverse working environment.
- This position will involve defense work. However, we will never build drones that can harm people.
Hit me up direct at { sturner at darkhive dot com }
(Not with the company) My observation is that ROS 2 is the go-to solution for new projects. Established co's have sunk cost in proprietary stacks and tend to spread the "ROS is slow and bulky" meme, but the reality is a sane, seasoned stack that has thousands of contributors offers a time-to-market advantage that is worth more than saving a bit on hardware. Most of the really time-critical code (actuator firmware, video capture etc) is done by C/C++ libraries anyway.
It's parallel to the general situation where Python is at the top of the stack in spite of being dog-ass slow.
ROS2 is also not ROS. I believe they fixed some of the mistakes. I'm not sure about catkin though, it's the worst build system I've experienced. And it sure likes to integrate itself everywhere, making it really hard to move away from it.
The only big change which isn't just churn is changing the messaging system to DDS. My problems with ROS1 aren't related to the implementation details of the messaging system. It's the build system, the difficulty of adding tests, the difficulty of writing deterministic systems with their messages being inserted in the middle of everything, needing to put everything through their messages, and in the end the fact that I can't use ROS for a system and non-ROS for another system and have them work together nicely.
You seem knowledgeable about this stuff — do you work in this space? I’m trying to do some feature extraction from multiple global shutters right now on limited hardware and its a learning experience!
Darkhive.com | United States | REMOTE | Full-time | $150-220k + early shares, employee #10-15
Darkhive is creating a palm-sized, autonomous drone optimized to demonstrate the capabilities of our software stack. The integrated software and hardware solution are key to realizing the full potential of the product in our target markets.
If ROS2, sUAS, GTSAM, SolidWorks, and aerospace are things you're familiar with or interested in, let's chat. Shipped a consumer product in volume? Let's chat. Maybe you're working an another UAS/Drone company that's larger and want to flex your brain to have more direct impact and ownership of your creations. Apply below and mention HackerNews.
Let's make drones more affordable and accessible to public safety, humanitarian, and defense applications.
- Position can be 100% remote (US Only). Relocation to San Antonio, TX preferred.
- We do in person meetings when needed. We are just 10 people.
- We're a startup, but we have funding and contracts. We're all mid-career so we want a sustainable, mature, diverse working environment.
- This position will involve defense work. However, we will never build drones that can harm people.
Darkhive.com | United States | REMOTE | Full-time | $150-220k + early shares, employee #5-10
Public safety and military robotics systems are unintegrated with existing equipment and applications. They require too much specialized training, making them incredibly niche and significantly inhibiting broad adoption.
Darkhive is creating a palm-sized, autonomous drone optimized to demonstrate the capabilities of our software stack. The integrated software and hardware solution are key to realizing the full potential of the product in our target markets.
If ROS2, DDS, sUAS, GTSAM, SolidWorks, and aerospace are things you're familiar with or interested in, let's chat. Maybe you're working an another UAS/Drone company that's larger and want to flex your brain to have more direct impact and ownership of your creations. Apply below and mention HackerNews.
We're looking for early hires to build out our SLAM stack, optimize the airframe, and build novel autonomous solutions. Our robots are designed and built in the USA. Let's make drones more affordable and accessible to public safety, humanitarian, and defense applications.
- Position can be 100% remote (US Only). Relocation to San Antonio, TX preferred.
- We do in person meetings when needed. We are just 10 people.
- We're a startup, but we have funding and contracts. We're all mid-career so we want a sustainable, mature, diverse working environment.
- This position will involve defense work. However, we will never build drones that can harm people.
I did software work for most of my career and am a huge fan of remote and hybrid arrangements.
But having started a hardware robotics company… it does seem more difficult to run the team remotely. When we get together in person, everything moves faster. We have more equipment. EEs can fix things on the spot. Perception engineers can perceive their algorithm’s behavior in front of their face. We can mitigate safety concerns with proper protections.
I wish we could all be remote, but I think the team does need a component of hybrid to pool resources and move faster. I only say this as there’s a lot of “C-suite bad they just want to micromanage” and I feel like it’s not that simple for many businesses.
Don't feel bad. Working/developing with real hardware seems to me a legitimate case where wfh might not work.
This does not mean that the constraints you have apply to everyone and most places that are forcing RTO are actually optimizing for other things like CRE.
Not every job can reasonably be done remotely. Working with specialized and expensive hardware is an area that seems to me to have a reasonable expectation that people work in the office, at least when they need to test on actual hardware.
Also why I'm avoiding such fields personally, despite thinking they're pretty neat.
Several more out there still from the initial bust.
Regulations make it challenging for beyond visual line of sight (BVLOS) autonomous delivery, particularly in the US with the FAA. Also, it's a hard problem in autonomy to nail it every time despite advancements. Less than 1% failure is still potentially catastrophic when you're carrying a few pounds of lithium batteries above people's backyards.
Love the hardware design and the no-nonsense "get it done" with continuous drone operations in Europe. Things for us to aspire towards as another drone (non-delivery) company.
Email in bio and would love to chat!