Exo ⓧ @exo_tech_

We Build Across AI, Blockchain & Product Engineering exotechnologies.xyz/?utm_medium=so… Joined May 2021

Tweets

1K
Followers

34K
Following

568
Likes

2K

Exo ⓧ @exo_tech_

3 weeks ago

Teach our team OR pitch something if it's really subtle and interesting :)

We've been doing lunch and learns @exo_tech_ where anyone from the team can present on a topic they're interested in. Anyone external interested in presenting to our team? Could be any topic you find interesting. Some topics we've covered: - Hazelnut farming - Bordeaux wine

5 0 20 1K 1

1 0 3 496 0

View Details

Exo ⓧ @exo_tech_

a month ago

DeFi Dev Corp holds over 2 million SOL. Not a fund. A publicly traded company on Nasdaq (DFDV) built entirely around accumulating and compounding Solana. We talked to their COO @TheOtherParker_ about what that actually looks like from the inside. @defidevcorp

1 5 21 3K 0

View Details

Exo ⓧ @exo_tech_

a month ago

Traditional finance is hiring blockchain engineers faster than crypto companies right now. Fidelity, BlackRock, JPMorgan, Baillie Gifford. All building on-chain infrastructure. The skillset they need: engineers who understand both TradFi compliance and DeFi architecture. That intersection is tiny. Demand is massive.

1 0 8 736 4

View Details

Exo ⓧ @exo_tech_

a month ago

8/ The compounding effect: a 30% reduction in CU usage across a protocol doing 100K transactions per day saves real money in priority fees. At scale, optimization is not a nice-to-have. It is a cost structure decision.

0 0 3 417 1

View Details

Exo ⓧ @exo_tech_

a month ago

7/ Preflight simulation: always simulate before submitting. It catches errors before they cost you fees and lets you fine-tune your compute budget based on actual usage.

1 0 4 444 0

View Details

Exo ⓧ @exo_tech_

a month ago

1/ Solana transaction optimization is not about writing clever code. It is about understanding how the runtime executes your instructions and eliminating the work it does not need to do. Here is a technical breakdown.

2 2 30 2K 18

View Details

Exo ⓧ @exo_tech_

a month ago

The embedded team model works because context is the most expensive thing in engineering. Our engineers join your Slack, your standups, your repo. They do not hand off code. They are in the architecture decisions. Clients stay because the model works. Not because of a contract. That is trust.

0 0 1 474 0

View Details

Exo ⓧ @exo_tech_

a month ago

sRFC 40: ERC-4626 for Solana. One standard interface for tokenized vaults. Deposits, withdrawals, share accounting. We submitted it to the Solana Foundation because building custom vault integrations for every protocol is not sustainable. Standards compound. Custom code does not.

0 0 4 450 0

View Details

Exo ⓧ @exo_tech_

a month ago

We're sunsetting Exo's AI auditing tool. We built it, shipped it, and learned a ton from the teams who tried it. The biggest takeaway: in smart contract security, the stamp of approval matters more than the tooling. That requires a full auditing practice, not just software. The AI and agentic tech behind it isn't going away. It's now part of how our engineering team works internally, helping us review code faster and more thoroughly for the teams we build with. More products coming. This was the first, not the last. Thanks to everyone who tested it and gave feedback.

1 0 7 517 0

View Details

Exo ⓧ @exo_tech_

2 months ago

7/ The teams that get these patterns right early save months of rework later. MCP is infrastructure. Treat it with the same rigor you give your REST APIs and database schemas. Reach out to talk: exotechnologies.xyz/contact

0 0 1 208 0

View Details

Exo ⓧ @exo_tech_

2 months ago

6/ Pattern 5: Versioned Schemas. Your MCP server will evolve. Version your tool schemas from day one. Breaking changes to tool interfaces break every AI integration that depends on them. Treat your MCP schema like a public API contract.

1 0 1 237 0

View Details

Exo ⓧ @exo_tech_

2 months ago

1/ MCP servers are becoming critical infrastructure for enterprise software. But most teams are building them wrong. Here are 5 patterns we have seen work across production deployments.