Pinned Tweet
Ok, here it is. This is game-changing, and what we are about to build and release on top of the @gnusai Operating System.
drive.google.com/file/d/1iuJ…
Ambitious? Maybe... But with our OS already built, IMHO, a 3-6 week project. With embedded proven blockchain technology.
Humanity, you are welcome.
@elonmusk, @PeterDiamandis "A day before something is a breakthrough, It's a crazy idea", @JeffPeoples, @beffjezos, @XFreeze, @tetsuoai, @BetterCallMedhi, @EricJorgenson, @gailalfaratx, @cboyack, @MarioNawfal, @Natortelli, @SawyerMerritt,@michaelnicollsx, @MrBeast, @Cobratate, @SawyerMerritt, @grok, @rohanpaul_ai, @Chad_Hurley, @stevechen, @sama, @brycent, @Ronald_vanLoon
So true, why we are building something unique.
The GNUS Cognitive App is a system that continuously models, challenges, and improves an individual’s thinking.
Built on top of @GnusAi a distributed, verifiable compute network that executes and coordinates arbitrary workloads—including AI, rendering, and cognitive processes—while continuously optimizing cost, performance, and correctness.
Per the FT, Anthropic would need about $602B a year in revenue to earn a 20% return on roughly $1T in compute commitments. That's more than Apple or Alphabet brings in.
There's a cheaper path. Billions of devices sit idle most of the day, and open models like DeepSeek V4.1 Flash and Tencent's Hy4-Flash already run on desktop and gaming hardware.
GNUS.ai turns that spare compute into AI capacity and pays the people who own it. The compute already exists. We don't need to build a trillion dollars of new data centers.
teapaw.org/unusual_whales/status/…
Anthropic would need about $602 billion in annual revenue to earn a 20% return on its roughly $1 trillion in compute commitments, per FT
The math: a 20% return on $1 trillion is $200 billion in after-tax profit. At a 17% tax rate, that’s about $241 billion before tax, and at a 40% operating margin, it takes about $602 billion in revenue.
That’s more than Alphabet’s $471 billion or Apple’s $478 billion in annual revenue. It’s about two-thirds of the $908 billion earned by all 76 software and IT services companies in the S&P 1500 combined.
For comparison, Anthropic’s annualized revenue run rate hit $81 billion at the end of October. Goldman, the lead underwriter, projects revenue of more than $600 billion by 2030.
Kenneth Hurley retweeted
Replying to @Object_Zero_ @firesidealpha
Agreed, but the 2030s infra isn't more giant training data centers.
Once we hit close-enough ASI, the game flips to inference, and that runs on a swarm of the world's idle GPUs, not $100B concrete boxes.
That's what we're building at @gnusai. More training DCs is building for 2024.
Looking forward to speaking at the Robots and AI Holiday Meetup in Oakland on Nov 18. Thanks @orbis86_x! I'll be talking about running AI on spare compute in devices people already own. teapaw.org/orbis86_x/status/21223…
✨ We’re excited to welcome Kenneth Hurley (@SuperGeniusEth), CEO & Co-Founder of Genius Ventures Inc., as a speaker at the 𝗥𝗼𝗯𝗼𝘁𝘀 𝗮𝗻𝗱 𝗔𝗜 𝗛𝗼𝗹𝗶𝗱𝗮𝘆 𝗠𝗲𝗲𝘁𝘂𝗽 on November 18 in Oakland, California!
Kenneth brings extensive experience in technology leadership, engineering management, and building innovative technology solutions. As CEO & Co-Founder of Genius Ventures, he is focused on advancing decentralized computing and next-generation technology infrastructure through GNUS.AI.
With a background spanning technology strategy, platform development, and business leadership, Kenneth brings valuable insights into how emerging technologies are shaping the future of innovation.
At the meetup, connect with Kenneth and fellow founders, investors, and technology leaders for conversations around AI, robotics, and the next wave of technological breakthroughs.
📅 November 18, 2026 | 6:00 PM – 9:00 PM PST
📍 Oakland, California
🎟️ 𝗥𝗲𝗴𝗶𝘀𝘁𝗲𝗿 𝗡𝗼𝘄: luma.com/tfot-robotosaiholid…
#TheFutureOfTech #KennethHurley #GeniusVentures #Robotics #EmergingTech #ArtificialIntelligence #HolidayMeetup
Kenneth Hurley retweeted
a few months ago, i reached out to a few VC friends to raise a small round for the swarm harness. turns out most were more interested in how you build a business around it than what we were actually trying to build. or maybe i just spoke to the less technical ones lol.
anyway, i kept building. fast forward, engineers from 15+ fortune500 companies have signed up for ArkenOS. researchers from startups building their own autonomous labs are using it. we're seeing a couple billion tokens of research activity daily. there's clearly demand. but i'm deliberately not optimizing for enterprise saas economics or short term revenue goals just yet (3 more months).
i want to understand the underlying physics of swarms. collective entropy, emergent specialization, how agents explore the tails of a hypothesis distribution, how knowledge propagates through an ecology, and whether we can get superlinear or better returns on every token spent.
there's probably a very profitable enterprise software company somewhere along this path. but i'd much rather build the machinery that makes an entirely new kind of scientific intelligence possible. sometimes you have to ignore the VC advice and keep building. this is the era where it gets rewarded the most.
if we're right with this, the way humanity explores the realms of knowledge could change quite profoundly.
if you're someone who wants to try or help me, my DMs are open! feedback and help are much needed and appreciated 🙏
In the grand scheme of ASI, I think we're vastly underestimating how much the harness matters. Pi, Codex, Claude Code, Cursor, Droids, and others can all spawn subagents, but spawning more agents doesn't fundamentally change the architecture. It's still a loop delegating work to other loops.
Why is an agent the fundamental unit of intelligence? Why does a swarm need to be a tree of subagents, with a single root deciding what everyone does? We're taking a different path with Optimus-1. The harness should provide the conditions under which new forms of collective intelligence can emerge.
An ecology of agents as the harness itself. Here's a glimpse into our harness!
The most sustainable data center is the one you never build.
GNUS.ai runs AI on spare compute in devices people already own. That means fewer new chips, less cooling, and less water.
On Nov 3, voters rejected data centers in 40 of 41 local ballot measures. We don't need more of them. teapaw.org/SuperGeniusEth/status/…
Replying to @AutoDid95368396
In a roundabout way yes.
Using spare computing capacity in devices people already own can cut the need to manufacture more chips and build more data centers, along with their water, cooling, and construction demands.
Distributing workloads also spreads power demand across existing infrastructure.
The savings extend far beyond electricity. That's why we built GNUS.ai to be the most sustainable Al platform.
Kenneth Hurley retweeted
Replying to @AutoDid95368396
In a roundabout way yes.
Using spare computing capacity in devices people already own can cut the need to manufacture more chips and build more data centers, along with their water, cooling, and construction demands.
Distributing workloads also spreads power demand across existing infrastructure.
The savings extend far beyond electricity. That's why we built GNUS.ai to be the most sustainable Al platform.
Kenneth Hurley retweeted
What actually happened: a capacity upgrade knocked servers offline, their load piled onto the rest, and those failed too. Outlook, Teams and Copilot went down worldwide for over a day.
That's what centralized failure looks like. Spread the work across millions of devices and no single change can take everything down.
Seems Microsoft likes the idea too: use spare computing power for AI and reward device owners.
My patent: US 11,451,393 B2 (2022).
Microsoft’s related patent: US 12,792,414 B2 (2026).
We’ve built this at GNUS.ai already.
instagram.com/reel/DfrcJpnN4…
What actually happened: a capacity upgrade knocked servers offline, their load piled onto the rest, and those failed too. Outlook, Teams and Copilot went down worldwide for over a day.
That's what centralized failure looks like. Spread the work across millions of devices and no single change can take everything down.
Kenneth Hurley retweeted
Replying to @sunilvallath @matthuang
But what if the token cost is 1000x less? I predict in the end the cost will actually be lower. Because distribution of costs to lower priced electricity and cheaper hardware already works.
The real constraint on swarm diffusion isn’t model quality or node count. It’s whether the orchestration surface stays invisible.
The architectures that actually remove friction treat the swarm as a drop-in execution substrate rather than a new system you have to adopt.
Existing agents, IDEs, internal tools, and workflows keep speaking ordinary OpenAI-shaped requests. A thin router turns those into signed jobs on a peer-to-peer queue; local execution is the default and the system escalates to multi-node only when complexity, risk, or value justifies it.
Developers never touch node registration, capability matching, CRDT queues, reputation scoring, or verification contracts. They change a base URL and an API key.
Once that translation layer exists, the “swarm rider” skill collapses to the part Huang already identified as scarce: defining clear reward functions and evals, then getting out of the way. The coordination tax that keeps conventional companies from becoming swarm-native largely disappears because the wires never appear in their process or their tooling.
Token spend can then go power-law through the long tail instead of remaining concentrated among the current set of swarm-native teams.
Swarm riders and Swarm-native companies
AI-use even among highly AI-native people is power-law distributed.
There seems to be a new class of people — call them “swarm riders” — who are capable of consistently commanding large swarms of agents to productive ends.
This is a new kind of managerial skill, analogous to the skill it takes to manage large organizations of people, but wholly different in nature and of course more technical.
Based on the swarm riders we know (including many at @Tempo and @Paradigm), these people are capable of productive token spend that is 1000x-10000x higher than median software engineers, and 100x higher than top 5% Claude Code users.
The art of swarm riding is different from software engineering and more akin to machine learning, where practitioners grapple with black box systems: defining reward functions and evals, transmuting non-verifiable problems into verifiable problems, running experiments, etc.
Swarm-native companies are a new class of company founded by and populated by swarm riders. These companies are organized around swarms in a way that conventional companies will struggle to replicate, as the cost of internal human coordination is sand in the gears of an effective swarm. Effective swarm-native companies will define the right reward functions and evals, remove functional and bureaucratic barriers, and then let the swarms cook.
We’ve gotten to see a few swarm-native companies up close — reshaping financial rails @Tempo, building a new gaming platform @Spawn, cyber swarms at Cathedral Systems — and it’s clear these teams wield a significant advantage over pre-AI teams.
Some questions I’ve been wondering about:
1) How fast will swarm riding diffuse through the economy? The implications for token spend could be significant.
2) Can swarms be effectively adopted by existing companies or will it generally track the formation of new swarm-native companies?
3) What becomes scarce as swarm capabilities themselves commoditize (through frontier models becoming better at swarming, swarm capabilities built into harnesses, etc)? Cultures and teams that can define reward functions well and otherwise get out of the way?
Wow, slowly but they are finally getting the tech I've built starting in 2016. @GnusAi
Agent swarms have the potential to drive yet another leap in output from AI. The impact will be a clear separation between teams that adopt this pattern of work and those that don’t.
The progression has roughly been: chat gave us quick turn AI responses that let us work a bit faster, but we were still rate limited by our interactions; then agents that can go off and complete entire tasks for us, accelerating hours or days of work into minutes; to now swarms of agents that are goal seeking and will find any route to the successful completion of a project, where there’s no inherent limit to the productivity gains.
We can imagine a future where, for any task in which value can be created within a self-contained digital environment, swarms will be able to complete the work faster than any other approach.
Essentially any area where value creation is currently bound by the amount of research you can do, the amount of decisions made with a set of software tools, the amount of data processed or consumed, and the amount of code (or code adjacent) that can be generated.
This means swarms will play a big role in game development, coding, trading, life sciences research, IT automation, cyber, math breakthroughs, animation, and plenty of other areas of work that have these characteristics.
There’s a lot of opportunity to build the harnesses, applied AI products, and infrastructure for how these agents operate. And most importantly there’s a ton of opportunity for those that can wire up work in this way. It’s going to be an entire new field of FDE work available.
Kenneth Hurley retweeted
Replying to @christophersaum
Guess how many of these requests actually respond? Zero.
Unpopular opinion: AI IPOs could be a bailout for VCs funded by retail investors.
US VC investors have put in $276B more than they’ve gotten back since 2022.
When these AI giants finally list, are you buying the future or paying for someone else’s exit?
pitchbook.brightspotcdn.com/…
Good read between the lines article about this.
pitchbook.com/news/reports/q…
1
42
Kenneth Hurley retweeted
So basically greater fool theory?
Someone pays more for a tiny slice, everyone marks up their shares, and suddenly they’re all richer.
Just need the next guy to believe it’s “real.”










