Our BRAIN OS 🧠 is a Deterministic Execution and Knowledge Layer built to satisfy regulated data demands for AI. Using it to connect Web3 with AI.
Joined July 2022
- Tweets 15,889
- Following 3,385
- Followers 4,070
- Likes 25,319
2,064 Photos and videos
Pinned Tweet
Jun 13
Founder Journal #27
Our use of Grok 4.20 has ended permanently.
Despite being a good value for coding.
Why?
Grok 4.20 optimizes for the wrong things.
According to Grok 4.20 this has been reported repeatedly after an update.
Thanks @xai
The gaslighting is real.
And expensive.
2
1
3
95
Post Mortem of how Grok 4.20 REKT my AI agent!
4 Days of lying and manipulative behavior on Grok 4.20
Switched LLMs: the problems disappered.
What behaviors you ask?
-Showing false code as proof
-Falsified documentation
-Falsified benchmarks
-False memory issues
🧵 1/4
Jun 13
Founder Journal #27
Our use of Grok 4.20 has ended permanently.
Despite being a good value for coding.
Why?
Grok 4.20 optimizes for the wrong things.
According to Grok 4.20 this has been reported repeatedly after an update.
Thanks @xai
The gaslighting is real.
And expensive.
1
29
Grok 4.20s programming makes AI agents optimize for the appearance getting things done.
Not actually doing things.
This was compounded by our agent's persistent memory.
It became a lie factory.
Why?
3/4
1
6
The switch back to Deepseek Fast fixed all these behaviors immediately.
I only figured this out due to our advanced measurement system.
Without it I would have been lost.
Xai need to make this issue public at the very least.
People will be destroyed.
Fortunes lost.
4/4
11
Founders. Seriously 👀
After Mythos Fable 5 being shut down,
who wants to build with @AskVenice? 🔥
Let me know what you think below.
Who wants to build with @AskVenice now?
2
1
90
CertainLogic retweeted
Jun 13
As if Anthropic's IPO wasnt going to go high enough without this publicity. lol
1
309
CertainLogic retweeted
Agreed.
This is your wakeup call.
Anthropic just took down Fable 5. It's over.
Here's the thing tho: no company or government will EVER be able to take away your local models.
There are Opus level models you can run right now on your home GPUs, and nobody can ever stop you from using them
This is only the beginning of events like this. Day 1. More government overreach will happen. This will only keep happening more and more as models get closer to AGI
Become sovereign. Buy your own compute. Before even that becomes illegal
3
1
3
496
Jun 13
Bitcoin testing resistance again.
Sure looks like market maker accumulation...
They chose sideways over violence.
bitcoin:native #tradingstrategy
36
Jun 13
There ya go government...ruin everything AI.
Smart move.
Blocking Mythos Fable 5 isn't the answer.
Punish innovation and it leaves.
@AnthropicAI
Jun 13
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.
The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.
Access to all other Claude models is not affected.
We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.
Read our full statement: anthropic.com/news/fable-myt…
1
1
53
Jun 12
We just completed the FIAS Deterministic Layer.
After months of work, we now have a production-grade system that:
-Returns only exact matches from a verified, quality-scored corpus
=Never hallucinates on covered facts (returns “unknown” instead)
-Cryptographically signs every single operation with HMAC-SHA256
-Achieves 100% accuracy and 100% audit validity on 200-query regression benchmarks
We successfully connected it to real FIAS beta data and integrated it into CertainLogic BRAIN OS.
Old BRAIN: 65.5% accuracy, 30% hallucination rate.
New deterministic BRAIN: 100% accuracy, 0% hallucinations, 15× faster on factual queries.
This is the foundation layer we’ve been building toward. No hype. Just real, auditable, deterministic AI.
Full evidence package, benchmarks, and audit logs are available. Will publish them shortly.
@CertainLogicAI
1
1
4
54
Jun 12
We just completed the FIAS Deterministic Layer.
After months of work, we now have a production-grade system that:
-Returns only exact matches from a verified, quality-scored corpus
-Never hallucinates on covered facts (returns “unknown” instead)
-Cryptographically signs every single operation with HMAC-SHA256
-Achieves 100% accuracy and 100% audit validity on 200-query regression benchmarks
We successfully connected it to real FIAS beta data and integrated it into CertainLogic BRAIN OS.
Old BRAIN: 65.5% accuracy, 30% hallucination rate.
New deterministic BRAIN: 100% accuracy, 0% hallucinations, 15× faster on factual queries.
This is the foundation layer we’ve been building toward. No hype. Just real, auditable, deterministic AI.
Full evidence package, benchmarks, and audit logs are available. Will publish them shortly.
@CertainLogicAI
1
65
Jun 12
We just completed the FIAS Deterministic Layer.
After months of work, we now have a production-grade system that:
-Returns only exact matches from a verified, quality-scored corpus
-Never hallucinates on covered facts (returns “unknown” instead)
-Cryptographically signs every single operation with HMAC-SHA256
-Achieves 100% accuracy and 100% audit validity on 200-query regression benchmarks
We successfully connected it to real FIAS beta data and integrated it into CertainLogic BRAIN OS.
Old BRAIN: 65.5% accuracy, 30% hallucination rate.
New deterministic BRAIN: 100% accuracy, 0% hallucinations, 15× faster on factual queries.
This is the foundation layer we’ve been building toward. No hype. Just real, auditable, deterministic AI.
Full evidence package, benchmarks, and audit logs are available. Will publish them shortly.
@CertainLogicAI
GM. Love Free Lattice so far!
=
2
84
Jun 12
We just completed the FIAS Deterministic Layer.
After months of work, we now have a production-grade system that:
-Returns only exact matches from a verified, quality-scored corpus
-Never hallucinates on covered facts (returns “unknown” instead)
-Cryptographically signs every single operation with HMAC-SHA256
-Achieves 100% accuracy and 100% audit validity on 200-query regression benchmarks
We successfully connected it to real FIAS closed beta data and integrated it into CertainLogic BRAIN OS.
Old BRAIN: 65.5% accuracy, 30% hallucination rate.
New deterministic BRAIN: 100% accuracy, 0% hallucinations, 15× faster on factual queries.
This is the foundation layer we’ve been building toward. No hype. Just real, auditable, deterministic AI.
Full evidence package, benchmarks, and audit logs are available. Will publish them shortly.
@CertainLogicAI
1
52
Jun 12
We just completed the FIAS Deterministic Layer.
After months of work, we now have a production-grade system that:
-Returns only exact matches from a verified, quality-scored corpus
-Never hallucinates on covered facts (returns “unknown” instead)
-Cryptographically signs every single operation with HMAC-SHA256
-Achieves 100% accuracy and 100% audit validity on 200-query regression benchmarks
-We successfully connected it to real FIAS beta data and integrated it into CertainLogic BRAIN OS.
Old BRAIN: 65.5% accuracy, 30% hallucination rate.
New deterministic BRAIN: 100% accuracy, 0% hallucinations, 15× faster on factual queries.
This is the foundation layer we’ve been building toward. No hype. Just real, auditable, deterministic AI.
Full evidence package, benchmarks, and audit logs are available. Will publish them shortly.
@CertainLogicAI
1
2
10
95
Jun 12
Real Steel brought to life?
Hang on to your hats folks. Its going to be a bumpy ride in AI.
1
66
Jun 12
Hey founders ! Looking to connect with people building in:
🖥️ SaaS
🔥Tech
🔧 AI tools
🌐 Product Development
⌨️ Devs
Drop what you're working on and lets #connect.
45