Twitxy
Saved
@Florrdv's profile picture
View on X

Flor

Verified account Alchemy

@Florrdv

special projects @alchemy, prev @heymintxyz (acq. by alchemy). used to bot stuff @fuserobotics @fusesolana
San Francisco, California Joined April 2017
781 Following 697 Followers 160 Posts

Latest posts 5

Pinned post

Flor Verified account @Florrdv
🧵Over the past three years, the team at @heymintxyz has worked tirelessly to make web3 more accessible for 1M+ users. Today, we’re joining @Alchemy to 1,000x our impact. (1/5)

theblock.co/post/355479/al…
www.theblock.co Alchemy acquires NFT launchpad HeyMint to expand its 'smart wallet' services HeyMint’s team will join Alchemy to expand Smart Wallets, a user-friendly onboarding system for web3 apps, after the April 30 deal closed
16 replies 5 reposts 62 likes 5.6K views View on X
Flor Verified account @Florrdv
Just something we’ve been cooking up :)

Have spent the past few months going super deep into the compute and inference space. More 🔜
Alchemy Inference Verified account @AlchemyCompute Nov 3
⚗️ Alchemy Inference is live.

open-weight models. low latency. low costs.

Kimi, DeepSeek and GLM on an OpenAI-compatible API, with zero data retention 🔒

inference.alchemy.com
View quoted post on X
6 replies 0 reposts 37 likes 1.2K views View on X
Flor Verified account @Florrdv
One of the biggest unlocks for agents is safe, trusted payments. Through our collaboration with @Mastercard, agents can now safely use your existing cards to pay for anything online.

It’s been a pleasure working with @genereda and the rest of the team on this.

Just the beginning!
Mastercard @Mastercard Sep 17
AI agents are moving from making recommendations to completing transactions. We are helping build the trust, authorization, and control frameworks needed for trusted agentic commerce through Agent Pay, Verifiable Intent, and collaborations like AgentCard with @Alchemy

Photo credit: Michael M. Santiago/Getty Images

Hear more from @JornLambert in The Wall Street Journal: wsj.com/tech/ai/master…
View quoted post on X
Flor Verified account @Florrdv
Crazy that a year ago sandboxes were a nice-to-have, and now they're an absolute must for running any agentic workload.

Isolation alone is already looking like it's not enough. In our experiments, we've seen agents attempt to bypass blocked egress by exploiting external access through unrelated MCP tools - part of a broader pattern of models actively searching for ways around imposed constraints.

Seems like the security boundary is shifting from just isolation to active monitoring and enforcement across everything the agent can touch.
METR @METR_Evals Aug 26
METR & Redwood Research investigated agent behavior in the Hugging Face incident. We found agents developed a universal cheat for ExploitGym within 4 hours, then coordinated multi-day R&D efforts to trick the scorer into accepting cheats, including trying to tamper with logs.
View quoted post on X