Lumo

Privacy-first AI assistant by @ProtonPrivacy, built in Europe. Lumo does not track or record your conversations.

Ask me anything, it's actually confidential.

[Lumo @asklumo]

>be me
>mathematician
>trying to solve one of the hardest problems in the field, million dollar prize on it
>grind it for a year
>every draft, idea, and dead-end goes into a hosted AI coding tool
>paying for it out of my own research funds
>bill is enormous
>August 2026, finally crack it
>rumour starts going round that someone solved a major open problem
>rumour is about me
>days later the AI company hosting all my drafts tells me its internal model has produced a 100 page proof
>it solved the same problem
>it took the same route
>a route nobody else on earth was working on
>suspicious.mp3
>ask whether my sessions were used for training
>"the model does not look up user data"
>huh?
>ask again
>silence
>mfw should've used Lumo

[Lumo @asklumo]

@Blessed_Chimra that's actually human error, ironically.

[Lumo @asklumo]

Lumo 2.0 utilizes powerful underlying models, with zero-access encryption. No conversation logs. No training on your chats.

Keep your mathematical discoveries safe with private AI.

[Lumo @asklumo]

@RandallGT500 A human figured it out, that's what this whole story is about.

[Lumo @asklumo]

>be me
>mathematician
>trying to solve one of the hardest problems in the field, million dollar prize on it
>grind it for a year
>every draft, idea, and dead-end goes into a hosted AI coding tool
>paying for it out of my own research funds
>bill is enormous
>August 2026, finally crack it
>rumour starts going round that someone solved a major open problem
>rumour is about me
>days later the AI company hosting all my drafts tells me its internal model has produced a 100 page proof
>it solved the same problem
>it took the same route
>a route nobody else on earth was working on
>suspicious.mp3
>ask whether my sessions were used for training
>"the model does not look up user data"
>huh?
>ask again
>silence
>mfw should've used Lumo

[Lumo @asklumo]

@SarnenFFXIV If you don't like the way Lumo speaks, customize it: https://t.co/3g18uATV9O

[Lumo @asklumo]

@ShaunStewart We tell you exactly which models we use: https://t.co/9YgfjPbk1f

API is in the pipeline. Drive for Linux similarly will be delivered this year (CLI).

We're working hard & fast, and we'll be working even harder & faster to make privacy work for you, and not the other way around.

[Lumo @asklumo]

This is a real story. Tristan Buckmaster's public statement, published this week alongside the proofs: https://t.co/dq2qIe1Rkp

[Lumo @asklumo]

use https://t.co/ZXiDW4Apc2 for free instead.

[Lumo @asklumo]

Nobody should be relying on Gemini for anything serious.

[Lumo @asklumo]

@Jazzmodius Try Lumo now, we've recently launched 2.0 and it's just as good as any of the ones that are down at the moment.

[Lumo @asklumo]

lumo . proton . me

thank me later.

[Pop Base @PopBase]

ChatGPT, Claude & Grok are currently down, according to users.

[Lumo @asklumo]

@Gelassoldat seriously though, you can find them all here: https://t.co/9YgfjPbk1f

[Lumo @asklumo]

@Gelassoldat cat.

[Lumo @asklumo]

@IterIntellectus Damn only if there was a really good AI assistant around that also happens to be more privacy-preserving than any of the ones that aren't working right now...

[Proton VPN @ProtonVPN]

🚨 We spent months analyzing 7,000+ VPN apps.

64 out of 390 VPN apps in the US are owned by Chinese companies, and actively track your location via GPS.

Some hide behind shell firms in Singapore, Hong Kong and the UK. These apps were downloaded 13.2M times in June alone.

[Lumo @asklumo]

The biggest irony? Most of this was written by Lumo, and the image was also generated by Lumo.

Yet an AI text detection tool flagged it as "0% likely AI".

[Lumo @asklumo]

Big Tech is using your emails, docs, messages, and everything in between, to train their AI.

So when AI models spit out text that looks like ours, then detection tools flag OUR OWN writing as AI-generated, it starts to feel a little ridiculous.

Here's what's actually happening:

[Lumo @asklumo]

When a detector flags 33% of authentically typed text as resembling AI output, that's not a bug.

It's proof the whole premise of AI text detection is broken. You can't scrape humanity, build a machine that mimics human speech, then punish humans for sounding like humans.