Co-Founder & CTO @get_hydrahost NVIDIA Cloud Partner | Miami Tech | b/acc | Building distributed AI infrastructure & compute markets
Miami, FL
Joined March 2014
- Tweets8.7K
- Following2.1K
- Followers2.3K
- Likes89.6K
Pinned Tweet
Entire Hydra team is excited to support El Salvador in its Sovereign compute initiative!
Bitcoin + AI Accelerators is the new playbook.
🇸🇻 BREAKING: EL SALVADOR MAKES HISTORY WITH NVIDIA B300 CHIP ORDER
President Bukele just secured the WORLD'S FIRST SOVEREIGN quote for cutting-edge B300 chips following a meeting yesterday with Aaron Ginn, CEO and Co-founder of Hydra Host!
These aren't just any chips - they're the most advanced artificial intelligence processors on the planet, and they're headed straight to El Salvador's National AI Lab.
🔥 All eyes on San Salvador - the future isn't just coming, it's being BUILT right now.
That's the inherent problem with Anthropic's paternalistic stance on blocking "offensive" capabilities.
Offense is the best defense. You're either sitting around waiting for officially disclosed CVEs and patches, until which you are just completely helpless. Or you can pen test the hell out of your own systems and find the exploits first.
I’ve heard several OAI people say there are unfixable bugs in hardware that are exploitable by AI models directly speaking machine code so even if we could solve software that’s not enough. We need defense to move faster than offense if that’s even possible we may just have a new normal that requires adapting to insecure software (bank transfers require going into a branch and taking a dna test 🤣😬)
Ariel Deschapell (b/acc) retweeted
The last thing you see before compute becomes a commodity @NotASithLord @andrawesbahou
Willing compute markets into reality at GTC Berlin
@ComputeDesk x @get_hydrahost
If the AI Infrastructure market has a collapse, it will be because we can't make GPUs efficient enough fast enough. Not the opposite.
Few.
Hello @FoundationHQ
Ariel Deschapell (b/acc) retweeted
unpopular opinion:
personal assistants are really not the most compelling usecase for cloud VMs sandboxes
he is right you can run them on nothing. the same way a real life personal assistant doesn't need a HPC cluster, just a phone + Acer laptop
agents hillclimbing in science, arts, engineering, is what will have a usecase for unlimited CPU compute
and those will be trillions of agents, running hundreds of trillions of vCPU, because the unsupervised post-economic value there is uncapped
assistants are for current human civilization, HPC clusters are for long horizon kardashev climbing
Takes enormous contextual knowledge and intellectual clarity to be able to consistently do the obvious thing.
Then those GPUs become even more valuable because you can squeeze a lot more out of them.
There seems to be an implicit and pervasive assumption that there is some ceiling to the amount of useful and intelligent work to be done. Once we hit that that’s it, we don’t need any more. That’s nonsense.
We’re on a non-stop freight train to Kardashev 1 so buckle up.
Every neocloud should be watching this.
What happens when the useful compute you can squeeze out of a GPU improves faster than the hardware itself?
Billions in GPU-backed financing depend on assumptions about future rental rates and residual values. Software breakthroughs could make those assumptions obsolete overnight.
Ariel Deschapell (b/acc) retweeted
Great moment at @NVIDIAGTC Berlin!
Hydra Host’s Head of Sovereign AI, Carl Meacham, had the opportunity to connect with @nvidia founder and CEO @JensenHuang.
As an NVIDIA Cloud Partner, Hydra Host is proud of our growing relationship and shared commitment to expanding access to AI compute worldwide.
We’re helping build the next generation of distributed AI infrastructure, enabling countries and businesses not just to consume AI, but to develop and deploy it on their own terms.
#HydraHost #NVIDIA #NVIDIAGTC #CloudPartner #AIInfrastructure
It’s starting.
See you on the other side.
It's ironic and pretty funny that everyone assumed that the very first thing LLMs would be really good at was obviously just writing, and it turns out they are actually terrible at doing it well.
It's incredibly obvious when something is LLM written, the reader's brain shuts off, and it's next to useless. Software development is by far much more useful.
Hydra Host has been working on the practical engineering problems surrounding compute markets and cross-vendor fungibility going back to our founding in 2021, far longer than anyone else in the space.
Today, it's exciting to have submitted comments to the CFTC on compute derivatives and contracts on behalf of the entire Hydra Host team reflecting our in depth experience on the subject.
For mature compute markets to take shape, one critical challenge is to that participants need be able to differentiate and group compute *quality*. This impacts physical delivery as much as it impacts derivatives and credit in the industry.
Check out the CFTC comment: regulations.gov/comment/CFTC…
Ariel Deschapell (b/acc) retweeted
Lots of VCs kindof sound like this: "Your basketball team doesn't have a moat, any healthy human can throw the ball into the basket thingy. "
People who have not run companies before don't understand how much the team matters more than some imagined structural moat.
Apple’s culture may knee jerk not like this, but at the end of the day it benefits them. The more flexible the hardware and the more compatible it is with different stacks, the higher the utility.
The higher the utility, the higher the residual value and secondary market for the hardware. That pushes demand for new units up.
Some interior state should just green light 30+GW of nuclear power, and an “industrial zone” with special provisions to fast track all permits for the associated build outs and infra for the power and data centers. Normal taxes apply and the State’s own investment funds GT mandatory inclusion for any project it wants.
Ariel Deschapell (b/acc) retweeted
We’re hiring three Network Automation Engineers: one in the U.S., one in Singapore, and one in the UAE.
• U.S., Singapore and UAE
• Full-time
• Automate networks across many vendors: SONiC, Arista, Juniper, Cisco and more
• Python or Go, Ansible, Terraform, BGP
Apply via the link in the comments.
#HydraHost #Hiring #NetworkAutomation #AIInfrastructure
If that holds true, then it’s not really a moat at all.
Apply this and you’ll realize the only moat at all is network effects.
Ariel Deschapell (b/acc) retweeted
When markets are growing quickly, moats don't really matter. It doesn't make sense for leaders to invest in new spaces. And tons of whitespace gets created.
When markets slow down, most moats can't save you anyway.
Don't worry too much about moats.
Ariel Deschapell (b/acc) retweeted
Have any of the labs tried to do in history what they've done in math, like digitizing the relevant documents, having lots of historians in the loop, and throwing a few billion tokens at a question like "why did the industrial revolution happen when and where it did?"