yelling at people to use more AI co-founder @merge_api

San Francisco
Joined September 2009
A few months ago I spoke at the AI Engineer World's Fair in SF on why your company needs a context graph. Connecting a few MCP servers won't tell your agent which customers were upset last week. Live API lookups can't search across that much data, so you need a synced copy for queries alongside live calls for simple lookups. Thanks @aiDotEngineer for having me. youtube.com/watch?v=cSz7aL2n…
4
5
51
1,466
MCP this, API that. It doesn’t matter. I asked a complex question and saw 133 calls in 41 seconds, with rate limits on almost every one. Physics, not protocols, is the bottleneck. It's the reason context layers are all the rage.
2
42
594
Gil Feig retweeted
Replying to @merge_api
@merge_api for Workforce is now live on the Microsoft Marketplace! 🔪🔪🔪 marketplace.microsoft.com/en…
1
3
27
491
Slack on the iPhone Duo is unreal
3
1
25
1,293
Gil Feig retweeted
Just onboarded one of the fastest-growing AI companies in the world onto @merge_api. We've made a lot of product bets over the past year, and it's incredible to see our team's hard work continue to pay off. ✨️
2
7
44
1,218
First official evals ran on our Context Layer V2 product coming soon... Vs. using MCP alone: - accuracy gains of ~40% - latency reduction by ~30% - cost savings of nearly 50% It's insane what a tiger team can accomplish
1
2
18
556
Non-flying cars and MCP are both dead
9
20
1,121
Phone cases are the AI slop of the physical products world
3
16
553
world’s most obvious answer clickbait
1
15
559
Reauth is what actually kills most integration migrations. Auto-link skips it: hand us a credential you already have, Merge creates the linked account directly, no consent screen. merge.dev
2
1
43
847
It’s amazing using AI products and instinctively knowing how they all work since we’ve all obsessively followed each AI improvement since ChatGPT dropped
2
15
690
fish can’t breathe on land gill issue
splattered ink on parchment quill issue
2
12
950
Gil Feig retweeted
Which model uses MCP tools best? We tested 6 models on the same 30 multi-step tasks, using real MCP servers in Claude Code. Frontier models like Opus 5.5 led on completion and speed but open models like MiMo cost 87% less per attempt. How much is Opus’s lead actually worth? 🧵
11
7
5
58
3,693
In China, robots in factories were supposed to increase economic output, but they displaced workers, and thus income and consumption decreased. Now their humans AND robots are unemployed. Is AI going to follow a similar story for knowledge workers?
7
413
The hardest part of putting an agent in production is giving it access without handing it the keys to everything. That work just got @Merge_api named one of @FastCompany's 7 next big things in foundational AI for 2026. Today, 1,700+ customers send 1B+ requests a day through Merge. Thank you Fast Company for the recognition! fastcompany.com/91612044/fou…
5
46
719
Gil Feig retweeted
Merge has been named one of @FastCompany's 7 Next Big Things in Foundational AI for 2026! Agents are only as good as the data they can reach. Merge is the layer that connects them to tools like @Salesforce, @Workday, and @GitHub, while keeping access secure. Check it out: fastcompany.com/91612044/fou…
7
2
38
1,434
We're looking for an AI pilled PM to join the team and work directly with me on owning our newest initiatives. Apply if you're perpetually on X, curious about the latest tools, and want to build without barriers for some of the world's largest enterprises and AI companies.
37
12
1
348
29,014
DM or reply and let's chat
8
18
3,038