American AI dominance @squaretower_
San Francisco
Joined December 2020
- Tweets282
- Following996
- Followers1.1K
- Likes1.9K
Mackay G Grant retweeted
i find it really odd (and uninspired in new literature) that the canonical human-element sought to be reserved for humans only vs. machines in sci-fi is emotion. and that emotion is the source of all human fallibility, but somehow drip-replicating it in machines improves their performance.
but has this actually panned out in practice? often, emotion is one of the first complex things intelligent machines have learned to imitate, and that their makers have actively tried to imbue.
the device that differentiates between replicants and humans in blade runner (1982) seeks emotion in the subject. and in blade runner 2049, controlled emotion makes the replicants more submissive somehow.
in pantheon, selectively adding memories back to an uploaded intelligence improves creative thinking/task adherence.
the butlerian jihad, wherever it has appeared in literature, is often a more subtle interpretation of this (erewhon, dune, even lord of the rings) that the more "human" a machine is, the more we serve them, and the more singularly evil.
Mackay G Grant retweeted
With @StorkOracle we are able to bring critical datasets to 72+ organizations, enabling Squaretower’s compute data to power new products that underwrite hundreds of billions of dollars in growing compute risk.
Compute is the most critical commodity of the next century. More than ever, we need transparent data and tools to understand how this commodity will shape and form economies of the future. This starts with Squaretower’s proprietary datasets, benchmarks, and research.
If you’re a team interested in data access — whether for understanding compute pricing or for new products — reach out.
Bringing GPU compute pricing on-chain.
@squaretower_ is building the financial infrastructure for the compute economy, aggregating GPU rental pricing across cloud providers and normalizing fragmented configurations into standardized, comparable metrics.
Now, Squaretower joins Stork as a data publisher, delivering high-fidelity GPU pricing directly to on-chain applications, starting with an NVIDIA H100 index.
Foreshadows well prediction markets becoming increasingly cybernetic. Each topic broken down into granular real time outcomes directly tied to individual financial underwriting, brings “the information war” into everyone’s pocket
POLYGLOBE JUST GOT ANOTHER MASSIVE UPGRADE 🌐
DeepStateLive’s Ukraine war map is now draped directly onto the globe, live and synced with @Polymarket .
Shaded control zones, shifting frontlines, Russian/Belarusian unit icons and direction of attack arrows all sitting under your markets and OSINT in near real time.
These Ukraine geo markets confuse people constantly. is it the city limits, the oblast, “enter” vs “capture all of”?
Now, when you hover a market, we draw the exact area of operation it resolves on and spell out the rule in plain English.
No more rule debates, no more guessing what actually settles.
The only way to monitor the situation.
Pricing the AI bubble; compute finance/markets are still early "wild west" - we won't find the top unless we can price risk correctly & there's no authoritative reference for any of this yet...
- Counterparties (colo, OEM, orchestrator, offtaker, lender, insurer, rails) demand proper risk transfer tools, not just term sheets or vanilla bilateral contracts
- Price/credit lens: reference curves/surfaces for actual traded $/GPU-hr (spot + term) + counterparty risk & credit / history of operators (exact data center specs, contract stickiness, etc.)
Most structures look like modest, tax-juiced income plays v.s. infinite money printers - expectations aren't aligned on duration, mkt value over time, capacity, and speculative activity. The right financial markets will be a massive expectation alignment force on all of this
We’ve spent the last couple of months deep in financing and coordinating our own B300 cluster deployment. Suffice to say buying, financing, deploying and monetizing GPUs is a wild process, so we figured we’d outsource our findings with ideas on improving this market. DMs open if you’re working on any of this:
1. First important to address that compute is product led not asset led. Consumption experience is the product, so buying bare metal is only part of the equation of bring a cluster to market (also why 1GPUhr does not = 1GPUhr). Orchestration and networking are critical processes to maintain on an ongoing basis. H/t to the start ups doing this as part of the monetization (selling the compute via contracts or marketplace) process- @fluidstack , @sfcompute - we are interested in meeting all players in this space in addition to 1. DCIMs optimizing workloads and networking (investors in @Aravolta_X25 ) 2. applications that enable further homogeneity of compute
2. Financing compute is a tough business mechanically and financially. Mechanically, there are several operational processes to underwrite and numerous counterparties involved (your colo, your OEM, orchestration and networking support, your offtake customer, your financing partner, your insurance carrier, whoever you use for financial rails). On these...
3. In underwriting offtake partners, look through into credibility is critical. @SemiAnalysis_ is leading the charge with ClusterMax performance data on neocloud and marketplace performance which is non trivially important for underwriting your ultimate customer - but there is further room to expand on this (credit ratings to complement performace ratings, actual market DATA on traded $/gpu hour for spot and term vs. prop indices of hypothetical quotes)
4. Underwriting your colo: there’s no real market to compare pricing and performance of colocation datacenter providers. This is a largely OTC, relationship driven process of knowing who to call and what your monthly rent/kw should like as well as what upfront costs should look like. Further, colocation billing and payments systems are archaic. We invested in gnomos (@iamminesoc) for this reason. Someone build a marketplace for colocation!
5. Financing partners: we are fans of what @USD_AI_ is building, they’re one of the only lenders who don’t require sale leaseback for those seeking upfront depreciation benefits in the US, with a novel mechanism to secure liens onchain. More of this - compute economy capital flows belong onchain.
6. Procurement: working with a strong channel partner is essential, buying a cluster involves customization of your BOM for your use case, negotiation, and securing credit and delivery terms. Outside of networking and orchestration, there’s technical work required in designing a cluster fit for your customer or use case. Monetizing with multiple customers vs 1 customer or for training vs inference present different designs in your networking, memory and control planes. Tons to unpack here, probably in its own report.
7. We modeled returns for GPU investing exhaustively for different tax, off take, and financing scenarios. Ultimately, this is a modest income generating investment with attractive near term tax benefits in exchange for longer term tax liabilities. There is demand for this at scale to be clear and we are doing our part to fill it. But care around assumptions on taxes, utilization etc. is paramount in not getting carried away in return expectations. Given that this is a tricky financial puzzle to solve, we are looking to support companies that bring more tooling to this process, some ideas above but a ton of white space to run here,
8. TAX. Last but not least, tax can be the widowmaker or the deal maker here. You need to properly account for and optimize your upfront depreciation (active/passive - there are ways to do each), your recapture, your upfront sales or use tax, and your ongoing income and b&o. Region matters, entity structure matters, etc etc
if you made it this far, DM with ideas/what you're working on, + a bonus meme for you:
many good founders blindly trapped chasing the bottom of this chart — most “hyped” firms are a porn brained capital wrapper. real risk capital doesn’t hide behind MBAs/analysts/content but just effectively supercharges asymmetric capital<>industry coordination
when actually the “shitty little apartment” is an 8br house in pac heights and everyone is a dropout backed by YC or a16z.
This Post is from an account that no longer exists. Learn more
there is an unconventional relationship style that young people disregard—i call this “journeyship”—you don’t have to keep up the relationship like every other, rather you find a deep initial connection and focus on following up spontaneously, every 6 months, 1 year, or even 3 years. life is long, there are many characters—it’s beautiful to check in on someone’s journey over a decades long period. these can be the most distinctly memorable types of relationships looking back on life.
there is a severe context issue w/ startup advice (esp in certain "accelerators"); you get blueprints on how to act which are essentially like getting handed chocolate coated catturds; the hard truth is that the only way to learn hard things is to do them your own way first
midjourney crushing it with opening up "full control panel" for prompting - also feels like a great edge over gpt in the sense that its both for creatives and thinking creatively with ai.
strength and weakness in society are cyclical—we build lasting structures not just to endure, but to remind us the value in creating what endures.
POV: having 20 back to back vc meetings and a call with your mom in the middle—but the your mom asks the toughest questions.
wisdom from Dostoyevsky—if you were put on trial now to be executed, what are your regrets? are you really living?
