Estamos en Venado Tuerto, en @rventurinosa, representante oficial de @JohnDeere, junto a @KamayVc. Recorriendo lo que se viene en agro.
Desde WakeUp Labs estamos sumandole valor a una industria que es piedra fundamental de Argentina.
Thanks for the shoutout!
Operator orchestrates what happens when AI already writes to production: the request, how it ran, who looked, what went live.
Whitelist open: [email protected], subject Operator.
Straight Outta LATAM: this week across the network.
@eldoradoio added new networks for sending USDT and USDC transfers.
https://t.co/dvbRbhOqhu
@MercadoLibre in 2025, 51% of SMEs selling on MeLi reached a city they hadn't before. 2 of every 3 shipments crossed a provincial border.
https://t.co/UQXMCgsrQQ
@wakeuplabs introduced Operator. Whitelist is open for early access.
https://t.co/pJpbkdrB70
@meta_pool the first cohort of its validator program wrapped with 55 nodes on NEAR Protocol and over 5 million NEAR staked.
https://t.co/P3BOlmoBKJ
@VitriumCapital Legal & Compliance Manager Brenda López on how the firm grew, scale and structure changed, the entrepreneurial spirit stayed.
https://t.co/Up4JuEdmPM
Para eso armamos Operator. Nos cambió la forma de trabajar internamente y ahora lo estamos implementando con clientes.
Si tenés este problema, metete en la whitelist y probalo antes que todos. DM o [email protected]
https://t.co/3qW1y9oxQe
Cuando la IA manda algo directo a producción, sea un humano promptando o un agente autónomo con permiso, ¿lo que salió está bien?
La pregunta ya no es si podés shipear. Es cómo reducís el daño: QA, security, legales, un humano, o nada, pero que quede decidido.
Most agent projects die in the API docs.
Auth flows, pagination, error strings you have to guess at.
We built the layer that skips that, and put the bounds in the interface definition instead of in a runbook nobody reads.
https://t.co/bR49xWlysN
Your agent can operate an @Uphold account now.
uphold-cli: one Rust binary, zero dependencies. Runs as a CLI or an MCP server.
Prices, trades, portfolios, accounts. Valid JSON, stable error categories.
Live execution requires an explicit confirmation flag.
New chapter.
We've partnered with @UpholdInc to build the financial infrastructure layer for AI agents.
Phase one is already shipped: a Rust binary that gives agents native, policy-bound access to Uphold's API.
Long-term collaboration. Read more ↓
https://t.co/bR49xWlysN
The hard part is not writing a dapp. It is getting it live on mainnet, with real users.
We did that 3 times on @arbitrum.
What we learned: 📷https://t.co/blPYaFR8fa
Built with @OffchainLabs Stylus, as a Mini App with @lemonapp_ar.
Demo: https://t.co/cUgkty8xXJ
Code: https://t.co/d3Wossr9jh
Post: https://t.co/M8ZhlQmPdW
cc @ArbitrumDevs
If Stylus is going to win DeFi, this is the kind of problem
it has to survive. What would you add to the whitelist?
One USDC deposit. Four live vaults on @arbitrum.
Vaulty (Stylus M3) splits across @aave, @Morpho, @0xfluid and @eulerfinance in a single tx.
The user owns the weights. Nobody else can move that money.
The next wave of blockchain users may not be humans.
Over the past 30 days, AI agents initiated 14.0M transfers through the x402 protocol.
Base led with 7.3M transfers, followed by Polygon at 5.6M.
Buscamos Technical Account Manager en @WakeUpLabs 💥
Perfil técnico + gestión de clientes + proyectos de sw. 100% remoto.
Vas a aprender una banda en muy poco tiempo.
👇
A blockchain cannot surprise itself.
Part 2 of our @arbitrum Stylus series: what has to hold true in the gap between requesting randomness and getting the answer back.
Locked solvency. Permissionless recovery. Zero admin knobs on the refund buffer.
https://t.co/ddjJs6qtkT
MCP solves discovery. It doesn't solve what happens after an agent finds the tool: policy bounds, failure modes, confirmation gates before touching production data.
We build that control layer for financial systems. DM us if you're building this.
Policy-controlled wallets are necessary.
But the real bottleneck is the layer between the agent and the core system: typed contracts, failure modes that agents can actually recover from, and confirmation gates on irreversible operations.
That’s the part we’ve been building in production.
As AI agents begin making decisions, accessing services, and initiating payments, the financial infrastructure behind them has to change too.
A new report from @Shoalresearch explores why the agentic economy may need programmable money, policy-controlled wallets, service discovery, and always-on settlement.
It also looks at how Circle’s broader platform, including USDC and Circle Agent Stack, fits that shift.
https://t.co/xr11AjB94K