In 2024, an employee at engineering firm Arup transferred $25.6 million after joining a video call where every other participant was a deepfake. Since then, AI has advanced significantly and is making impersonation increasingly convincing, causing challenges for financial institutions in particular. As autonomous agents begin handling more financial activity on behalf of people, establishing an added layer of trust and security behind a transaction or sensitive conversation is becoming an urgent need.
Introducing @world_id for banking and finance, a suite of technologies that helps institutions establish trust through privacy-preserving cryptographic Proofs. World ID for Zoom lets meeting participants confirm they are the same Orb-verified human connected to their account, while World ID for agents allows AI agents to carry Proofs delegated by people. Bain Capital, OpenExchange, Outtake, Pantera Capital, and VanEck Funds are initial design partners, bringing World ID into some of finance's most consequential interactions.
https://t.co/NakSyrdBtq
@joulsounet He mentioned they release everything under feature flags, have the agent watch logs during the rollout, and will auto rollback if it sees elevated errors
One of the coolest talks at Frontiers was on a fully AGI-pilled company. Amp runs with 21 co-founders. Would encourage a listen.
1. No PRs or code review
Code goes from push to prod in about 7 minutes, around 90 times a day. Agent threads are shared instead of PRs
2. No product roadmaps and one meeting a week
I thought this was fun... they write a news post first. If you can't write the launch post, you don't know what you're building.
3. No support team
An agent fixes the bug and emails the user back in under 15 minutes.
4. Never ration tokens
Rounding errors next to employee salaries
5. "Software is dead"
He thinks the best people will move into manufacturing and finance.
Could be one of the coolest marketplaces at scale.
Tools have meaningfully improved my own Hermes agent performance, but you run into issues with some being unmaintained, having potential security issues, and needing to stitch together a cheap and effective set of tools for the task.
Played around with this using ToolRouter, so very excited to see this develop!
INTRODUCING MERCATOR by @tempo.
Mercator gives agents a new effort level.
By giving an agent a search engine to the world's paid tools it empowers a small model to punch above its weight, while being cheaper than larger models.
$5 Mercator credits w/ your X/Github ⬇️
introducing World One
a 3 month residency for solo founders in SF
residents get everything they need so they can focus on building: a great place to work, meal service, and a $10k monthly stipend for anything else
come build the future of the internet powered by proof of human, you can learn more and apply at https://t.co/3ggTzp8f4k
Education still remains underdeveloped as part of the ai buildout. No one is thinking about it as thoughtfully as @rshanreddy and team. Have heard incredible success stories from parents and students alike using this product.
Congrats on the launch!
America is banning AI in schools.
China is using AI to create geniuses.
Introducing Aristotle: The AI tutor that solves America’s broken education system. https://t.co/hqPJN8cke0