To be honest there's really no secret sauce in there. It's primarily how to get started with agents, when to abandon your context and start anew, and advice on models, monitoring cost, and prompting. This is not to diminish the value of the information as it's good information written by great colleagues. I just wanted to note that most of the information can be obtained from the official AI provider documentation and blog posts from AI boosters like Thorsten Ball.
Medication is littered with warning labels but humans still use it to combat illness. Social media can harm mental health yet people still use it. Pick whatever other example you'd like.
There are things in life that have high risks of harm if misused yet people still use them because there are great benefits when carefully used. Being aware of the risks is the key to using something that can be harmful, safely.
We're working on publishing an updating pricing page to better communicate this to the public. Stay tuned!
Here's what I can knowingly share. We sell full-rack and half-rack SKUs today coming in at 1024/16TiB and 2048/32TiB of CPU cores and RAM respectively. Disk varies depending on the drive size. On top of the cost of the rack you'll pay an annual percentage for support and that's it. No software licensing or other weird fees.
Hey! The current generation rack is compatible with the current generation compute sleds, switch, and power shelves. The next-generation compute sleds will be compatible as well. Possibly the next-generation switch but I would have to defer to my colleagues for that answer.
We'll ultimately ship a next-generation rack to take advantage of changing constraints (e.g., physical form factor, power requirements) and we'll provide a mechanism for upgrading to that generation of rack as well. Likely, in the form of multi-rack connectivity to move workloads or perhaps even financial incentives to purchase the latest and greatest.
I don't work on the hardware side at Oxide so I don't have further details to share. Great question!
Similar to what others noted earlier I'm having trouble understanding exactly what you're trying to communicate here. I'll respond based to points I am clear on.
Selling a customer a contract for on-premises computing and giving them a fake metal box and SaaS is borderline unethical depending on the terms of said contract. I understand the sentiment of that point though. There are many reasons a customer chooses to own instead of rent. Legal requirements, financial incentives, and even control over performance to name a few.
On-premises computing was so good that the cloud providers packaged it up and sold it back to people at a premium that could only ever be rented. The finances of that model don't make sense to many businesses as they look to reignite their on-premises computing with the modernity of the cloud providers. That's where Oxide shines in my opinion- being able to have on-premises computing that combines the efficiencies of the hyper scalers with an API-driven approach to managing resources. We take that a step further by building hardware and software in-house for additional benefits such as power efficiency, control over the networking stack, additional telemetry, etc.
there's seriously nothing complicated about this. Define ownership. Who cares about owning a bunch of computers? If you can get the compute, and you pay up front for it once, and then you can sell "it" later, and it's cheaper than renting, what does it matter if the compute comes in the form of a physical box or if it is in the cloud? So these are all things that matter about ownership! To me, occupying a physical space is not important for ownership, wrt servers.
Nobody is saying anything about anything unethical... I am mocking the idea of needing the steel box, forget about the steel box.
It's just Amazon Reserved Instances, but with an indefinite period. Okay? Isn't that an attractive product?
Why am I talking about banks? Because maybe in Oxide's deck it says, "Amazon will NEVER do this. Amazon will NEVER sell indefinite reserved instances." Fine. Well a bank can simply pay spot prices and sell you an up front price, if you want. Okay? It's the same thing. It only matters what Amazon does when we're talking about $100m Series B, which is what this article is about! It's not about the technology.
> On-premises computing was so good that the cloud providers packaged it up and sold it back to people at a premium that could only ever be rented.
No... guys... AWS makes sense. It's not a premium "that could only ever be rented." There are a ton of much cheaper cloud providers. Amazon just happens to be selling the Rolls Royce of clouds. They have a ridiculous margin. Figma makes more profit for AWS each year than it will ever make for itself in its entire lifetime. "that could only ever be rented" is simply not true, they can afford to make all sorts of innovative pricing models, reserved instances being one of them.
Oxide just hasn't had to compete with "99 Year AWS Reserved Instances." But absolutely, positively, utterly nothing stops them from offering that. They already give you a massive, MASSIVE discount for 3 year reservations.
That said, obviously not having to deal with human beings managing hardware is valuable. It's the same shit as the difference between "AI" meaning a computer and overseas workforces. They might produce the same outputs for the same cost, but think deeply about yourself: how much are you willing to pay to deal with a computer instead of an IT tech? To avoid phone calls? To avoid doing things that might be faster, but are in person?
You are casually transitioning between something that is pretty well understood through the history of online commercial computing (i.e. 1970s): 1) owned hardware versus leased/timeshared/rented hardware or services and 2) concepts of financial instruments that many readers here are probably not intimately familiar with.
If I needed a generator to run a remote mining operation, and you just told me to just buy energy futures instead, we'd be having a silly discussion. Whether it makes sense for me to rent or buy the generator has more to do with governments, [,tax ]laws, and risks that ultimately manifest as cashflow decisions. You have some valid thread you are pulling on for what are the economics of general purpose compute and to whom, but your argument needs a lot more care to carefully define and make your case and why it is okay to dismiss the outlier cases for instance.
> If I needed a generator to run a remote mining operation, and you just told me to just buy energy futures instead, we'd be having a silly discussion.
Exactly. Just because they are similar in some senses doesn't mean that they're fungible. Generator manufacturers still have a business even though you can purchase energy futures.
It would be like if you needed to buy a generator versus, I will give you a giant cable that has all the same economics as buying a generator. In that case yeah, if I am giving you all the same economics as buying your own generator, but cheaper, you would be stupid not to take my deal. It’s got nothing to do with energy futures. My suggestion is to copy what I am saying to a chatbot, and ask it what “looks like a duck, quacks like a duck” means, and what’s going on here.
"giant cable" -- this kind of thing does obviously happen, rural electrification is heavily subsidized by governments (see the rest of my comment..). But it is also not a realistic option in plenty of cases as such and therefore industrial companies like Cat and Cummins are happy to sell multi-million dollar generators. If your approach to the subject is to copy it into a chatbot it might explain why this discussion is ultimately specious, because a casual and even jocular approach to this is not really adequate to make any point.
I'm not part of Oxide, but I think you're assuming everyone is okay with not controlling the hardware they run on.
There is plenty of addressable market where a public cloud or even a colocation of hardware doesn't make sense for whatever bespoke reason.
I don't think their target customer is startups, and that's okay. You likely aren't their target customer. But they've identified a customer profile, and want to provide a hardware + software experience that hasn't been available to that customer before.
It's not clear to me that the counter-party risk is _remotely_ similar between the two offerings. Someone reselling indefinite access to AWS could relatively easily set things up to go "poof" after a few years, whereas if it's your box, that's not possible. Now, critical components of the box could croak, and someone could still be screwed, but even there, you likely have more options for support than just paying Amazon's ransom.
Oxide is selling a technical solution to technical customers with ownership needs or desires for security, regulatory, or other reasons. Access to the complete stack source code is another. I know a bajillion of these customers and they all have very, very deep pocket books.
They are specifically going after features in a way that no other vendor is, with an extreme care of execution of their crafts at the highest level.
The problems oxide is solving for these customers is something Amazon has shown no interest in. Could they? Sure, they could do anything they put their pocket books to doing, but they haven't.
Loved reading your experience here. Thank you for posting it. I've written about the value of an artifact in the past when people pushed back against the Oxide materials saying they are a lot of work for no guarantee. When I first applied to Oxide I was also rejected and the materials process taught me a ton about myself and changed the way I viewed job searching and my work. I shifted course and increased my skills and next time I applied I got an offer. There's power in the critical thinking and writing the materials force out of us.
Well, its the assymetry of wanting a 10 year long documented CV with various orthogonal points in your career, versus actually having a 30 minute call.
Unlike an actual interview, which is equal time investment, this 20 page paper gets the commentary and result of "no". "No" what? You can ask an interviewer about concerns, and discussion points. This email from no-mail@ is just nothing.
And its not the sting of rejection. I've been turned down, and I too have turned down. But its the mechanistic, dispassionate, legalistic response after months of a "No". And not even a 'What we're looking for is.... '
It's tough on both sides. I understand the disdain of receiving a "no" after putting in hours or days of effort on the materials. Candidates are welcome to ask for feedback on their application but must understand that Oxide is limited in what we can say due to legalities. Hiring is a tricky balance on the legal front. I also understand that it's impractical to give every candidate a synchronous screening call just to confirm whether they should continue applying. Not only for time's sake but also bias. Is it enough for 1 person to do a screening call and decide someone's fate?
No hiring process will ever be perfect but at least, as the experience mentioned previously touched on, the candidate is left with an artifact that they can then use in future applications. Candidates walk away having learned something about themselves. Open roles are also limited, applying for such roles is also optional, and it's up to each candidate to decide how much time and effort they wish to put into the materials.
We're humans here at Oxide too, and we're doing the best we can to ensure the hiring process is fair and humane as well.
The more I compare Material 3 Expressive to Liquid Glass the more I'm excited to switch back to a Pixel. I'm a fan of the use of color, motions, and different shapes versus transparency, minimal contrast, and little color.
I'm on a iPhone 13 Pro Max right now that's still doing well but starting to show battery capacity age. Plus, it's the only non-USB-C device I own so I'd be happy to get rid of it.
This WWDC was the nail in the coffin for me. Apple has lost the plot and is back to self indulgent projects. Picked up a Pixel again and glad I don't have to deal with liquid glass in the future.
I feel similarly. This iPhone is the only Apple device I own outside of a shared Apple TV. I'm not in the Apple ecosystem so moving back to Pixel for me would be easy.
I'm going to wait for the Pixel 10 to see what specs it has. I believe you're right though it's a bit behind on other manufacturers in terms of hardware. It would still be a jump for me coming from an iPhone 13.
To be honest there's really no secret sauce in there. It's primarily how to get started with agents, when to abandon your context and start anew, and advice on models, monitoring cost, and prompting. This is not to diminish the value of the information as it's good information written by great colleagues. I just wanted to note that most of the information can be obtained from the official AI provider documentation and blog posts from AI boosters like Thorsten Ball.