"A handful of companies will literally own the infrastructure of digital labor."| Credit: Hameem Sarwar /CCN
Share
Key Takeaways
The AI race has moved beyond chatbots to systems that execute real tasks. Large Action Models, not open-source LLM clones, will define the next phase of disruption.
The real money in AI comes from automation infrastructure, not consumer chat subscriptions. Enterprise workflow integration drives revenue and long-term dominance.
Centralized Action Models risk concentrating control over digital labor in a handful of corporations. Whoever owns the automation layer captures the value it creates.
Web3’s ownership economy cannot survive if it fails to compete in AI. Without community-owned automation systems, the promise of distributed value breaks down.
We’re fighting yesterday’s battle.
Right now, the decentralized AI community is proudly launching its tenth open-source large language model (LLM), while OpenAI and Anthropic are quietly building AI that can use software, fill spreadsheets, and run ad campaigns and practically every other computer-based task.
The real disruption happening in AI has nothing to do with better chatbots. Now, automation is taking over.
And if Web3 doesn’t pivot immediately, we’ll miss the automation wave entirely, leaving AI ownership completely in corporate hands.
Bet on ESports with These Partners
Sponsored
Disclosure
We sometimes use affiliate links in our content, when clicking on those we might receive a commission at no extra cost to you. By using this website you agree to our terms and conditions and privacy policy.
Why Building More LLMs Won’t Win the AI Race in 2026
The entire decentralized AI movement seems hypnotized by language models trained on publicly accessible language data.
Every week brings another “uncensored” or “privacy-preserving” LLM that does exactly what ChatGPT does, just worse and more expensive. We’re competing in a race that OpenAI already won two years ago.
Meanwhile, the real action is happening elsewhere.
Google is developing AI that can navigate websites and complete tasks.
Microsoft is building agents that manipulate Excel and PowerPoint.
AI-driven automation will displace hundreds of millions of digital jobs in the next five years: customer service, data entry, basic coding, content creation, administrative tasks, you name it.
If these LAMs remain centralized, the profits from automated labor flow directly to Silicon Valley shareholders.
The same companies that disrupted industries and concentrated wealth are about to do it again, but faster and more comprehensively.
Every efficiency gain, every productivity boost, every eliminated job. All the value accrues to Big Tech’s bottom line.
Plus, centralized AI means centralized decisions about what gets automated, who has access, and how these systems operate.
For example, Google will decide which workflows its AI will execute.
Microsoft determines which jobs its agents can perform.
A handful of companies will literally own the infrastructure of digital labor.
For web3, this represents an existential threat to our entire narrative. We’ve spent years building the “ownership economy,” filled with Decentralized Finance (DeFi), non-fungible tokens (NFTs), and decentralized autonomous organizations (DAOs), all premised on distributing value to participants rather than platforms.
But if we can’t compete in AI, that narrative collapses. Whoever controls the models controls the future of work.
Building AI That Actually Works
The path forward requires abandoning our obsession with language models and focusing on action-based AI with community ownership.
First, we need models trained on real interaction data, like clicks, workflows, and task sequences. Not Wikipedia articles and Reddit posts, but actual recordings of people doing digital work. This data is far more valuable than text, which is why Big Tech guards it so carefully.
Users won’t contribute this valuable workflow data for free. They need real incentives, such as tokenized ownership in the models they help train.
Every contributed workflow, every recorded task sequence, every bit of expertise should translate into equity. Turn data labor into actual ownership stakes.
“The path forward requires abandoning our obsession with language models and focusing on action-based AI with community ownership.” | Image source: Sina Yamani
We should create marketplaces where experts can encode their skills into executable AI agents. Here are two examples:
A tax accountant could train an agent on their workflow and earn royalties whenever someone uses it.
A social media manager could package their expertise into an automation that businesses can onboard within minutes.
Expertise becomes a tradable, scalable asset rather than hourly labor. By contributing to the ecosystem, either through training, workflows, marketplace or distribution, users should earn a fractional stake in the entire ecosystem.
This is no longer a debate about model quality. It is a race to control the infrastructure of digital work.
The Window Is Closing Fast
Some teams get this. A few protocols are building decentralized compute for Action Models or creating infrastructure for on-chain AI agents. But they are exceptions in an ecosystem still obsessed with building ChatGPT wrappers.
Google’s Workspace AI handles routine administrative work.
Every month, they ship new capabilities while we debate decentralized open-source language models nobody uses.
The timeline is brutal.
The Last Chance to Compete in AI
The decentralized movement has maybe 18 months to pivot before the automation infrastructure gets locked down completely.
Once enterprises standardize on Microsoft’s action models, once Google’s agents become the default for digital tasks, the game is over. Network effects kick in. Switching costs become prohibitive.
We need to stop competing in the last war.
Open-source chatbots won’t save us when corporate AI agents are doing everyone’s jobs.
The decentralized movement must build the AI that matters, AI that is effective. And that AI must belong to the people whose work it replicates, not the platforms that profit from it.
Build action-based, community-owned AI now, or watch from the sidelines as Big Tech owns the automated future.
Disclaimer:
The information provided in this article is for informational purposes only. It is not intended to be, nor should it be construed as, financial advice. We do not make any warranties regarding the completeness, reliability, or accuracy of this information. All investments involve risk, and past performance does not guarantee future results. We recommend consulting a financial advisor before making any investment decisions.
About the Author
Sina Yamani
Sina Yamani is the CEO and founder of Action Model, where he leads a team of 40+ engineers, operators and marketers building the next generation of autonomous AI. With a background in blockchain and emerging technologies, Yamani founded Action Model to create a fairer, community-owned alternative to today’s monopolized AI landscape.
Action Model is pioneering Large Action Models (LAMs) AI systems that don’t just generate text, but can click, type, and complete real tasks online – while ensuring ownership and rewards flow back to the people who help train and improve them. Previously, Yamani built and exited a multi-award-winning fintech company after leaving Cardiff University’s Computer Science programme, and was named Young Entrepreneur of the Year in 2021. He has been an advisor and speaker in blockchain and AI since 2017.
