iamdev - building MnemeBrain

Building MnemeBrain

Belief memory for AI agents

Contradictions • Evidence • Revision

Think TCP/IP for agent memory

[iamdev - building MnemeBrain @Iamdev_ai]

@MrBeast dot

[iamdev - building MnemeBrain @Iamdev_ai]

Are you curious? Follow me for more..

[iamdev - building MnemeBrain @Iamdev_ai]

Coming soon: GenoMark.
Autoresearch at home.
ZK verification for every result.
Train locally, submit confidently, prove it cryptographically.
🧬⚡
@karpathy @ensue_ai @MinaFoundation @MinaDevelopers @MinaProtocol @minacryptocom

[iamdev - building MnemeBrain @Iamdev_ai]

@boxmining Building MnemeBrain — a belief memory layer for AI agents.

Instead of storing text like RAG systems, agents store beliefs with evidence, contradictions, and revision over time.

Think TCP/IP for agent memory.

https://t.co/IQ7ZzNvdnA

[iamdev - building MnemeBrain @Iamdev_ai]

@X4AES @BLUECOW009 We have benchmark this:

https://t.co/XqPkVP5on3

[iamdev - building MnemeBrain @Iamdev_ai]

@X4AES @BLUECOW009 Exactly. The hard part isn’t storing memories — it’s how new evidence reshapes old beliefs.

Once agents run for weeks/months, memory becomes a dynamic system:

new evidence → belief update → confidence shift → possible contradiction

That update rule is still an open problem

[iamdev - building MnemeBrain @Iamdev_ai]

@navtechai @BLUECOW009 You need a way to track:

• evidence sources
• timestamps
• confidence
• conflicting states

Otherwise the system either overwrites or averages the truth.

[iamdev - building MnemeBrain @Iamdev_ai]

@navtechai @BLUECOW009 The hard part isn’t compression, it’s belief consistency.

If two memories contradict:

“user is vegetarian”
“user ate steak”

Compression alone can’t represent that conflict.

[iamdev - building MnemeBrain @Iamdev_ai]

@lenny_enderle @karpathy Logs tell you what happened.
Beliefs tell you what you’ve learned.

That’s what we’re trying with MnemeBrain in autoresearch.

[iamdev - building MnemeBrain @Iamdev_ai]

@karpathy https://t.co/OsP0gIJAKL

[iamdev - building MnemeBrain @Iamdev_ai]

@karpathy Most agent memory today looks like this:

logs → embeddings → similarity search → context

But that pipeline can’t represent things like:

• contradictions
• confidence
• evidence over time

So instead of storing text…

MnemeBrain stores beliefs with evidence.

[iamdev - building MnemeBrain @Iamdev_ai]

@Starwatcher_vc Thanks

[iamdev - building MnemeBrain @Iamdev_ai]

@Starwatcher_vc We benchmarked this — most memory systems score 0% on contradiction detection.

👉 https://t.co/inTIuGbRBJ

[iamdev - building MnemeBrain @Iamdev_ai]

@Starwatcher_vc Better retrieval helps.

But even perfect retrieval doesn’t solve contradictions.

Agents eventually store:

“user is vegetarian”
“user ate steak”

Vector systems retrieve both.

They still can’t represent truth state.

That’s a different layer: belief maintenance.

[iamdev - building MnemeBrain @Iamdev_ai]

@sukh_saroy We benchmarked this.
Most agent memory systems score 0% on contradiction detection.

👉 https://t.co/inTIuGbRBJ

[iamdev - building MnemeBrain @Iamdev_ai]

@sukh_saroy Persistent memory is the easy part.

The real problem is contradictions.

If an agent stores:

“user is vegetarian”
“user ate steak”

Vector memory retrieves both.

Which one is true?

Most systems push the problem back to the LLM.

Agents need belief revision, not just storage.

[iamdev - building MnemeBrain @Iamdev_ai]

@nextxAG @intheworldofai So instead of:

old → deleted
new → stored

we get:
supporting evidence ↑
conflicting evidence ↑
truth state recomputed (TRUE / FALSE / BOTH / NEITHER)

This lets agents track evolving beliefs instead of flipping state.

[iamdev - building MnemeBrain @Iamdev_ai]

@nextxAG @intheworldofai Good question.

We avoid hard overrides.

Each belief keeps:
• evidence graph
• confidence score
• timestamps

New evidence triggers belief revision, not replacement.