Mentioning their startup’s name, right at the top, isn’t a great start for me. OP didn’t need to do that, it means nothing to me, and it doesn’t add anything other than advertising. A few other phrasing choices like “full technical details on our website” evoke a recruiter spiel a little bit, because of wrong time and place — this is a comments section, and to me the writeup is a bit too detailed and overly confident in how interested I am. If I care, I’ll ask, or I’ll go to your profile.
Sounds like OP was just excited; unfortunately, the practice of using HN to “organically” advertise startups (particularly through blog posts) is quite common nowadays, and I can’t help being sensitive to it as I feel it doesn’t help discourse. This post was relevant and interesting though; thanks for flagging it as such.
Why don't I see down arrows? There are no down arrows on stories. They appear on comments after users reach a certain karma threshold, but never on direct replies.
Also explained here:
https://github.com/minimaxir/hacker-news-undocumented/blob/m...
I just voted you up to get one more point closer!
More than anything, your comment just feels like you just copy-pasted it from the blurb you send to investors? For example you say:
> ... while we can deploy any compute that our customers want, we are starting with AMD instead of Nvidia...
> One of my goals has also been to get Dell ... to offer compute credits on our cluster. ... It becomes a developer flywheel...
I mean, OK? The second part in particular ("It becomes a developer flywheel") seems totally irrelevant to anyone except a potential investor. Why would I as a customer care about your product being a flywheel??
I can hardly speak for everyone on this website, but I know I'm here first and foremost because I love to learn (c.f. "If you had to reduce it to a sentence, the answer might be: anything that gratifies one's intellectual curiosity."). Even if I were working on a startup similar to yours, your comment doesn't really teach me anything ("we're building our own first cluster of 16x Dell chassis..." OK?)
Also, quite frankly, I just went to your website and it reads like it was generated by ChatGPT.
> In essence, Hot Aisle is not merely a cloud compute provider but a dedicated partner that accelerates the journey of businesses towards HPC advancements, ensuring they navigate the digital transformation landscape with assured resource scalability, enhanced security, and unwavering support.
The developer flywheel was in response to the mentions of credits in the Google announcement. It isn't my product that is the flywheel, it is the imho smart concept of attracting developers to a solution by giving them credits. "free drugs", if you will. This is, in my eyes, a big reason why Nvidia is so popular today and what made the cloud in general, so successful.
I'm trying to bring that concept to the AMD ecosystem. Previously, you could only get access to AMD MI (enterprise) class compute, if you had access to super computers like El Capitan and Frontier. I'd like to bring these things to the masses and a big part of that is lowering the barriers as far as possible.
It seems stretch to call the annoucement as any kind of validation for your startup, whatever your exact logic is here.
Your reply above makes no sense to me. NVidia isn't popular because of any "free drugs". Cloud did not become popular simply by giving free credit.
It's not at all clear what value your startup is brining to the world. El Capitan is 40MW compute. You can get that much computer from top cloud providers (with money of course) - their combined compute is in tens of GW range, estimate based on their renewable power portfolio. The barrier nowadays is roughly only money, and unless you have magic to lower the price of power and machines, you are not lowering any barrier vs top providers. If you have the magic sauce, it's not clear what that is, at least in the post.
Their startup isn't Google. To me, it's interesting to hear how startups are competing with Google.
We're serving the need for individuals and businesses that don't want to use Google in the first place ^. Those of us who feel that there should be alternative hardware available other than just Nvidia, or even cloud solutions. We give full bare metal access to the underlying systems. We work with customers to customize their stack to tailor things to their use case. We're more 1on1 niche and solving for the least common denominator, with best-in-class solutions. Eventually, we are going to run hardware that others wouldn't normally touch and/or we are going to work to get it deployed earlier than any large cloud can.
^ Just as a side note, I'm a long time huge fan and customer of GCP, so please don't take this as bashing Google at all.
But ok, best of luck!
The validation is clear to me, but this is my field to recognize that. My business is about building super computers and either renting them piecemeal or whole to people and businesses. This is exactly what Google is doing in this announcement, which I take as validation because I started working on this similar thing, about a year ago now.
Nvidia built software and hardware (s/h) before AI. Nvidia ensured that all of their s/h solutions were easily available to developers. AI recognized that the s/h was useful and took advantage of that. An example of "free drugs" were to make large gifts of the s/h to colleges [0].
I'm not saying that cloud only got popular with free credit, but it definitely was a contributing factor (just an example: [1]) in the building of many startups.
The value of my startup is something I describe above in my original comment. You're dead wrong that the only barrier is money. It is the experience and relationships that we have in building, deploying and running large scale compute. Think of us as a consultancy for super computers. We also have the backing to fund the capex so that businesses don't have to put out millions up front, on rather finicky cutting edge hardware.
Not everyone needs 40MW of El Cap, all the time. Not everyone wants to deploy into a cloud, many want to have more control over where their compute and data is located. We work with Dell, AMD and data centers directly to build and deploy these systems. I won't talk about pricing other than to say that both companies are highly incentivized to work together to deploy as much compute as they can, and I'm the one that has joined with them to make it happen. I'd say that there are about 25-30 people involved with us, just to deploy our single first cluster. It is a massive amount of coordination.
It takes years of relationship building to even get your foot into the door on this. It is far more complicated than just racking boxes and we already have put the time and effort in to create the blueprint designs for best in class compute. We help companies that want this compute deployed yesterday, to speed up the whole process.
I'm sorry if that is not valuable to you personally, but it is to others.
[0] https://developer.nvidia.com/higher-education-and-research
My startup (Hot Aisle) is all about building, managing and deploying dedicated compute clusters for businesses. At the enterprise level of compute, there is a lot that goes into making this happen, so we are effectively the capex / opex for businesses that don't want to do this themselves, but want to have a lot more control over the compute they are running on.
The twist is that while we can deploy any compute that our customers want, we are starting with AMD instead of Nvidia. The goal is to work towards offering alternatives to a single provider of compute for all of AI.
You can't do this for others unless you also do it for yourself. As such, we're building our own first cluster of 16x Dell chassis with 128 MI300x GPUs deployed into a Tier 5 data center as our initial rollout. Full technical details are on our website. It has been a long road to get here and we hope to be online and available for rental at the end of this month.
One of my goals has also been to get Dell / AMD / Advizex (our var) to offer compute credits on our cluster. Those credits would then get turned around into future purchases to grow into more clusters. It becomes a developer flywheel... the more developers on the hardware, the more hardware needed, the more we buy. This is something unfamiliar to their existing models, so wish me luck in convincing them. Hopefully this announcement helps my story. =)
Edit: Getting downvoted. Would love to hear some dialog for why. I don't really consider this an advertisement, so apologies if you're clicking that button for that reason. I'm really just excited about learning about validation of my business model and explaining why.