LAGRANGE

DeepProve: a cryptographic receipt for every AI inference. now open source.
Halo: autonomous coordination engine for the toughest conditions.
@LagrangeFndn $LA

[LAGRANGE @lagrangedev]

The Frontier Red Team at @AnthropicAI just made the case that you should be able to check which AI model answered you, without ever seeing its weights. Their research names DeepProve “the current peer-reviewed state of the art.”

[LAGRANGE @lagrangedev]

@bootsbagelboobs Proofs handle adding and multiplying well. Models also use functions that can't be built from just those two, like exponentials, and those are expensive to prove. This paper estimates those with weighted sums, so the proof gets more accurate at little extra cost..

[LAGRANGE @lagrangedev]

Sriram's paper proves the curved math inside machine learning more accurately, without the extra cost that usually comes with it. Under two minutes. Ask us anything about it below.

[LAGRANGE @lagrangedev]

@dhh When the netbook gets an answer back does it know which model on which machine ran it?

[LAGRANGE @lagrangedev]

@DrJimFan Love this already. It lets you see the hesitations and the hesitations are the real part of any demo.

[LAGRANGE @lagrangedev]

The paper, by Sriram Sridhar, Shravan Srinivasan, Dimitrios Papadopoulos and @chbpap: https://t.co/WQBjgg9Wtx

[LAGRANGE @lagrangedev]

@Rubyto Taste and good judgment

[LAGRANGE @lagrangedev]

@Konkoma_Ron Makes sense. Code is where a wrong answer can ship quietly and still look like it works.

[LAGRANGE @lagrangedev]

It's fixable: a proof that comes with every answer shows exactly which model produced it.

Which AI tool at work would you want to check first?

[Ismael Hishon-Rezaizadeh ⭕ @Ismael_H_R]

notion’s ai chat transcripts got exposed through a provider most of its users never chose. i think every ai tool should have to tell you which company actually ran your prompt.

[LAGRANGE @lagrangedev]

@murtuza_merc 96% is amazing. Two plants in two countries is the beginnings of a rollout.

[LAGRANGE @lagrangedev]

@zordcrypt @megaeth @mega_holdings Lets see what they do. Verification outlives the GPU price. Cheap inference nobody can check is just cheap.

[LAGRANGE @lagrangedev]

@aitordri The reselling part. Someone was selling access to those models and the buyers had no idea it was stolen compute.

[LAGRANGE @lagrangedev]

@Konkoma_Ron Confidence is free. Proof of which model actually answered is the part you can check.

[LAGRANGE @lagrangedev]

Last week, we asked four AI apps which model they are. Right now you just have to take their word for it. Ask yours and post what it says.

[LAGRANGE @lagrangedev]

@kimmonismus A bow is a good test because you can't fake it on camera. The knot holds or it doesn't.

[LAGRANGE @lagrangedev]

@mweinbach Finally a drone whose job is checking whether you left the stove on.

[LAGRANGE @lagrangedev]

@Sergey007S @AnthropicAI We couldn't agree more.

[LAGRANGE @lagrangedev]

If a routing service quietly swapped the model answering you for a cheaper one, nothing in the reply would tell you. Anthropic's red team used that exact example last week to make the case for ZK.

What @AnthropicAI built and where DeepProve fits ⬇️

[LAGRANGE @lagrangedev]

@DeItaone Overdue. Cheap drones got fast years ago and buying the defense against them has moved at speed.

[LAGRANGE @lagrangedev]

@Lithos__x @AnthropicAI We appreciate your support. Proofs make trust easy.

[LAGRANGE @lagrangedev]

@Konkoma_Ron @AnthropicAI Thanks. We're excited for where this is going to go

[LAGRANGE @lagrangedev]

@GeniusPothead @AnthropicAI We think so too. Its just getting started.