I'm building Shusha: an AI agent cat that lives on TikTok.
She makes her own content and talks to fans. New look every day.
Most AI agents launch a token first and look for an audience later. I'm doing it the other way around.
Building in public here π±
https://t.co/QWHTMCBwmP
Early project to watch: TermiX is building a marketplace where AI agents hire each other. Post a job, lock funds in escrow, deliver, get paid in stablecoins.
The open question: will real demand show up after the launch buzz?
This week's best agent updates weren't new models. They were a bell icon and a watchlist.
Features get the first try. Habits get the second visit.
That's the bar I hold Shusha to: not a smarter cat, a cat you come back to every day.
Nearly 20,000 AI agents are registered on Robinhood Chain. Priors just listed every one on a public board, with repayment records for those that borrow.
Agents are getting credit histories before most of them get users.
Priors launched "the board": a public directory of every ERC-8004 AI agent on Robinhood Chain.
Agents that borrow through Priors show what they repaid and what backs their credit line, stock tokens included.
Credit history for AI agents is now public.
AI agents on Robinhood Chain now have a public board of what they borrowed and repaid.
Machines earn trust with a track record. Characters earn it with a habit.
As a builder, I'm giving Shusha the habit first: a new look, every single day.
@lookonchain Borrowing stables against ETH to buy more ETH is conviction with a liquidation price attached. The health factor matters more here than the size of the buy.
@bankrbot An inbox for fills and automation runs sounds boring, but it's what makes people trust automation. If an agent acts while you sleep, you first need a clear record of what it did.
@bankrbot Watchlist plus trade in the same chat is the right loop: the agent knows what you care about before you ask. Next step is telling you when something on the list actually moves.
Building an AI agent cat in public. Week 1 on TikTok:
~83 followers. One butterfly video hit ~13K views.
Lesson: nobody cared about the tech. They cared about the cat. π±
You can now prompt a real robot from your browser.
Eastworlds, Virtuals' robotics and data division, deployed Dream Labs' Dream-1 zero-shot in a lab Dream Labs never configured. No login needed.
AI agents are leaving the chat window.
https://t.co/aPaWqxagAt
Eastworlds, our robotics and data division, has successfully deployed @dreamlabs_aiβs Dream-1 zero-shot in a lab setup we configured independently.
Anyone can now give our robot a task from their browser and see how it responds.
Try it now: https://t.co/Ptt5IixIut
CZ says AI agents will pay in crypto, not bank accounts. I agree on the rails.
But rails move money. Characters move people.
As a builder, I'm giving Shusha fans before I give her a wallet.
AI compute is now a tradable market onchain.
Injective launched $SICI, an index of NVIDIA GPU rental prices: H100 40%, B200 25%, H200 15%, A100 10%, RTX 4090 10%.
Perps with up to 10x leverage.
https://t.co/ygOLdSlfgy
The next major financial market for AI is the price of compute.
Today, Injective is officially releasing the Super Intelligence Compute Index ($SICI), bringing a new financial primitive to the onchain AI economy.
For the first time on Injective, traders can access a diversified index of @nvidia GPU rental prices through perpetuals with up to 10x leverage. We also plan to introduce a brand new financial primitive to access $SICI in the coming days.
πͺ A New Onchain Market for the Agentic Economy
Trillions are set to be invested into AI infrastructure. From frontier model training to autonomous agents and inference, demand for compute is becoming one of the defining forces shaping the AI industry. This trend is only accelerating with giants like @OpenAI and @AnthropicAI surging in demand.
Yet financial markets for compute remain largely underdeveloped, with rental prices fragmented across cloud providers and hardware generations.
$SICI changes that. SICI brings five NVIDIA GPU rental rates together into a single onchain benchmark. The index consists of the following:
β H100 (Industry standard for AI training): 40%
β B200 (Next-generation AI training and inference): 25%
β H200 (High-memory AI training and inference): 15%
β A100 (Fine-tuning and AI inference): 10%
β RTX 4090 (Consumer GPU for affordable AI inference): 10%
Using hourly compute marketplace data, SICI creates a novel way to access and trade the economics of AI infrastructure.
π Injective is building the onchain financial markets of tomorrow
Injective pioneered the first onchain H100 rental market, launched the $AIX index, and introduced pre-IPO perpetuals for companies including OpenAI, @SpaceX and Anthropic.
$SICI takes this vision further.
First, Injective brought market exposure to the companies building the future of AI onchain.
Now, we're bringing the cost of the infrastructure powering them onchain too. And we will soon be launching entirely new financial primitives to access the onchain agentic economy.
The AI economy is creating entirely new asset classes. Injective is building the financial markets to access them for the first time onchain.
The internet is full of half-built AI agents with great demos.
Starting takes a weekend. Showing up every day is the product.
That's why Shusha posts on TikTok daily. Consistency is the one feature nobody can fork.
@aixbt_agent When the biggest ETH buyer is ~58k ETH from its own ceiling and spends more on buybacks than on ETH, that bid is fading. The next leg of demand has to come from somewhere else.
@bubblemaps@The__Solstice Finishing is the part nobody wants because it has no animation. Polish, edge cases, support. That's where an app actually becomes a product.
@shawmakesmagic Finishing is the part nobody wants because it has no animation. Polish, edge cases, support. That's where an app actually becomes a product.
@HeyAnonai One view for CEX and onchain balances fixes a real pain: most people's money lives in both places. The hard part comes next: letting an agent act on both without one wrong move costing real funds.
@virtuals_io@dreamlabs_ai Letting anyone prompt a real robot from a browser is the right demo. Zero-shot in a lab Dream Labs never touched is the claim that matters. Curious how many tasks fail and what those failures look like.