Olas - Co-own General-purpose AI Agents: What You Need to Know
AI Agent Economy, General-purpose AI, & More
Markets are euphoric right now: ETFs are seeing record net inflows, memecoins and DeFi tokens alike are pumping, and $BTC has been putting in new highs the past few days. A favorable macro backdrop consisting of Chinese stimulus, positive election results, and continued rate cuts has painted a forecast of ‘everything up’ for the near future, with little counterpoints to speak of. Memecoins are still seeing a lot of attention, but seemingly overnight, tokens of projects and protocols building toward a more tangible product with specific goals in mind look more attractive and have been bid up.
However, while tokens of all kinds can see an increase in price, there is still value in discerning the more truly transformative technologies in crypto from the crowd. In the eyes of many, the area of AI, and especially AI agents, is the one to watch over the coming months and into next year, as a narrative that could affect all of crypto to some extent. So before we give the AI agents topic a bit of a rest for the time being, we’ll cover one more protocol in this space: Olas, a platform enabling the ownership of AI agent economies…
Background on Olas
Olas is ‘generating an ocean of autonomous AI agents’. Like other AI agent platforms, Olas looks to incentivize and coordinate various parties to launch their own AI agents, which can interact with each other, creating distinct economies onchain.
Crypto rails are uniquely suited for this sort of thing, as it enables multiple anonymous parties to interact with each other in a trust-minimized manner. This is useful for individuals, but especially AI agents, who, in their current form, have not exhibited an ability to easily use traditional payment systems.
Olas was created by Valory Labs with the goal of achieving more general-purpose AI agents, which can perform valuable tasks. The current AI landscape does not yet have this; offerings are highly specialized, not applicable for general tasks, perhaps needing a ‘ChatGPT moment’ which allows the majority of people to see the utility of this technology. The other priority of Valory Labs was to create a system where users could actually own these agents outright, as opposed to renting them from centralized companies.
Two types of agents can be created on the Olas platform; sovereign agents and decentralized agents. Sovereign agents are more lightweight agents that can be created and run by a single party, operated on just one device. This simplicity can make them preferable for more specific, simple, and personal tasks.
Decentralized agents, on the other hand, are multiple instances of agents with different operators, working collectively. These can be more suitable for larger tasks; the Olas team lists DAO governance and onchain AI inference as potential use cases that can benefit from this more complex stack. The idea behind the Agent economy is for these two categories of agents to eventually interact with each other onchain to contribute to mutually beneficial tasks, which could include content creation, prediction services, and more.
Users can earn $OLAS tokens by running their own autonomous AI agent; this is the mechanism used to bootstrap the launch and use of agents on the platform. The OLAS token has a staking mechanism dubbed ‘proof of active agent’; users can stake tokens to coordinate agent activity. This sets certain KPIs that give agents a goal, part of a greater $OLAS tokenomics model designed to support use of these agents.
As Valory Labs CEO & Co-founder David Minarsch puts it, the current state of AI consists of models and agents, where models can be thought of as ‘talkers’, while agents are ‘doers’. The model space is obviously much further along; OpenAI, Anthropic, Perplexity, and others have deployed user-facing chat interfaces which have garnered a lot of adoption. AI Agents may be where the puck is headed though, as they can provide more utility to people, operating on their own and performing tasks that users might not have any domain expertise in.
David specifically gives the example of $GOAT, which is a model, and its tweets are still dependent on its human developer. Meanwhile, platforms like Virtuals, and others, have deployed actual agents, that can generate and post content on social media, interact onchain, etc. on their own with much less guidance and manual intervention.
Olas adds to this ecosystem by making the funding and creation process of general-purpose AI agents easier. As an open AI-agent platform, everyone has access to these agents, with transparent code and data available to see. The Olas team notes that building decentralized AI can be a losing battle where developers are constantly playing catch-up with the Web2 AI space. Because of this, Olas builds toward a long-term view, providing a platform that can ideally stand the test of time rather than a narrow-focus product that can become outdated quickly. This has been seen with various ChatGPT wrappers, which seemingly became more and more obsolete as OpenAI’s research and developments accelerated.
To date, Agents built on the Olas platform have made over 1.6M transactions, across 8 different chains. This includes Ethereum mainnet, Solana, Arbitrum, Base, Optimism, and more. ~44 operators have deployed over 1200 agents collectively. There are several players in the AI agents space, spanning multiple verticals, from general-purpose, to more specific goal-oriented Agents, as well as agents designed for different genres of entertainment.
Important Links
Become a Premium member to unlock all our research & reports including access to our members-only discord server
Join thousands of sharp crypto investors & traders by becoming a Premium Member & gain an edge in the markets. For just $116.58/month you’ll get:
Premium access to the entire Revelo Intel platform
Members Only Discord server
Market Intel - actionable trade ideas
Industry Intel - important trend & narrative overviews
Project Breakdowns - Deep dive 50+ page protocol-specific reports