LAGRANGE (@lagrangedev) / X

LAGRANGE

3,300 posts

LAGRANGE

@lagrangedev

Verifiable computation for AI, defense & enterprise. Halo: coordination engine for autonomous systems. DeepProve: prove every AI inference. @LagrangeFndn $LA

Ethereum

linktr.ee/lagrangedev

Joined May 2022

Following

100.3K

Followers

Pinned
LAGRANGE
@lagrangedev
Jun 24
A drone swarm is only a swarm until the network breaks. After that, it's just a lot of drones flying alone. Everyone's racing to build more autonomous systems. Almost no one's building the engine that keeps them working as one. Today that changes. Introducing Halo, the
3K
LAGRANGE
@lagrangedev
5h
This is a drone swarm in the conditions that break everything else: nodes dropping offline, comms jammed, GPS denied. With Halo, the swarm doesn't break. Halo keeps every drone on the same shared picture, reassigns the work the instant one drops, and reconfigures the swarm in
00:00
1.9K
LAGRANGE
@lagrangedev
5h
1 of 4 short films. The next three show what coordination makes possible once a swarm can survive contact with the real world. See how Halo works → lagrange.dev/defense
309
LAGRANGE reposted
LAGRANGE
@lagrangedev
17h
Comms jammed. GPS denied. For most drone swarms, the shared picture should fall apart. Halo keeps every drone in agreement straight through the kind of electronic warfare that breaks everything else.
1.9K
LAGRANGE
@lagrangedev
17h
Comms jammed. GPS denied. For most drone swarms, the shared picture should fall apart. Halo keeps every drone in agreement straight through the kind of electronic warfare that breaks everything else.
1.9K
LAGRANGE
@lagrangedev
17h
Halo is the coordination engine for autonomous systems. Built for the densest jamming on earth. One shared picture. No central control. See Halo in action → lagrange.dev/defense
279
LAGRANGE
@lagrangedev
Jun 24
Replying to @lagrangedev
When comms are jammed, GPS is denied, and drones are dropping offline, Halo keeps the swarm in agreement. Every node holds the same picture. Failure is caught without false alarms. Roles reassigned the instant one drops. The system reconfigures itself in real time. No central
529
LAGRANGE
@lagrangedev
Jun 24
Built for the teams fielding autonomy: defense primes, drone OEMs, robotics and autonomy teams, and program offices. See how Halo works →
Lagrange - Verifiable Autonomy for Defense
From lagrange.dev
429
LAGRANGE
@lagrangedev
Jun 22
AI incidents hit a record high last year. up 55%. model transparency fell for the first time. down 18pts. (Stanford AI Index 2026) more AI. more harm. less disclosure. "trust me bro" is not an incident report. would you ship an AI decision you can't prove?
2.5K
LAGRANGE
@lagrangedev
Jun 22
the receipts already exist. proving an AI actually ran the way it claims is exactly what DeepProve does.
DeepProve · The fastest AI proof system · Lagrange
From lagrange.dev
531
LAGRANGE
@lagrangedev
Jun 19
Lagrange Labs is now a part of @IBM’s Partner Plus Program.
1.5K
LAGRANGE
@lagrangedev
Jun 16
AI is quickly becoming infrastructure. It’s moving money, running enterprise systems, making decisions. And we’re still treating it like a tool you can double-check later. You can’t. Once AI is infrastructure, “probably correct” stops working. Full piece:
1.2K
LAGRANGE
@lagrangedev
Jun 15
71% of executives say they won’t scale AI without proof that it’s correct. Turns out “source: the model felt good about it” is not enterprise infrastructure. DeepProve gives every inference something enterprise AI needs: proof you can actually check. 👇 lagrange.dev/deepprove
1.1K
LAGRANGE
@lagrangedev
Jun 12
For decades, digital systems were built on trust. We trusted servers to run the right code. We trusted networks to relay accurate information. We trusted platforms to verify identity and content. In an AI-driven world, trust is no longer enough. We are moving from trust-based
1.7K
LAGRANGE reposted
Allora
@AlloraNetwork
Jun 10
🔴 AMA: Allora x @lagrangedev Thursday, June 11 · 3PM EST / 20:00 UTC Allora's Head of GTM @MichaelZski will sit down with @BrianNovell from Lagrange labs to break down how ZK coprocessors make AI verifiable, and what it takes to prove an Allora inference end to end. Got
3.4K
LAGRANGE reposted
nikkolasg
@nikkolasg1
Jun 3
Finally deepprove is open sourced 🎉 the repo is here github.com/Lagrange-Labs/… and you can run your own bench for gemma3 and gpt2 (and llama2 but not optimized). It's quite modular so adding another model is easy ! Reads safetensors, gguf, onnx !
LAGRANGE
@lagrangedev
Jun 3
AI doesn’t need to be a black box. So we opened ours. DeepProve is now open source. The fastest zkML in the world. Fork it. Run it. Break it. Build something weird. → lagrange.dev/blog/deepprove…
00:00
GitHub - Lagrange-Labs/deep-prove: Framework to prove inference of ML models blazingly fast
From github.com
2K
LAGRANGE
@lagrangedev
Jun 9
“60× faster zkML” sounds like something marketing would say. Unfortunately for marketing, the benchmarks are public. Fork it. Run them. Beat us if you can. → lagrange.dev/blog/deepprove…
1.1K