Kanishk Patel

Building Memory Layer for Physical Life

[Kanishk Patel @kanis_patel]

What is DuoMem? https://t.co/w3kgSgeWJu

[Kanishk Patel @kanis_patel]

What is Agents’ Last Exam? The benchmark where top agents drop from 82% to 25% https://t.co/nISkNtHoJP

[Kanishk Patel @kanis_patel]

Tool-call hacking: why your deep research agent cites sources it never read https://t.co/pRK1nHZhPj

[Kanishk Patel @kanis_patel]

What is the Darwin Gödel Machine? https://t.co/EhQnpa1RhS

[Kanishk Patel @kanis_patel]

https://t.co/8p5ecrzuVm

[Kanishk Patel @kanis_patel]

So I saw a lot of people using Fable 5 to built CAD modelling and since 2021 I have been wondering if models will become good enough to work and visualize in the 3D space...

Also one hot prompting Fable 5 to create a complete 3D game to use LLM's to build stuff was insane.. check it out https://t.co/7et2oaMrmk

[Kanishk Patel @kanis_patel]

@boardyai curious what Boardy Pro looks like, count me in

[Anthropic @AnthropicAI]

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: https://t.co/bwn0sximKZ

[Kanishk Patel @kanis_patel]

@the2ndfloorguy Well we know who is getting fired now! Lol

[Kanishk Patel @kanis_patel]

Why Your Agent Problem Is a Spine Problem, Not a Brain Problem https://t.co/6WUS2XsXOE

[Kanishk Patel @kanis_patel]

What Are Self-Evolving Agents? https://t.co/w6FLyFkTHj

[Kanishk Patel @kanis_patel]

Every Major Agent Benchmark Just Got Hacked. Here’s What Still Holds Up. https://t.co/wvE3IFOwVw

[Kanishk Patel @kanis_patel]

@rachelrapp I went to one of the Baseten, SF tech week event a couple years again while I was building my own startup. Impressive what they have achieved in a short period of time.

[Kanishk Patel @kanis_patel]

What Is the Lethal Trifecta? https://t.co/mnUxS2UqYF

[Kanishk Patel @kanis_patel]

Founders overcomplicate this, 5 steps:

1. Claude Code builds the MVP. @hotjar from day one.
2. AI does outreach. Real users in the door.
3. Claude watches Hotjar for friction, rage clicks, dead sessions.
4. Claude proposes improvements. You approve. It ships.
5. Talk to customers yourself. Agents drift without a human compass.

The loop runs while you sleep. You stay in the loop so it doesn't go off the rails.

[Kanishk Patel @kanis_patel]

Set Up Harness Engineering on Your Python Repo in 30 Minutes https://t.co/r8rtZRi07p

[Kanishk Patel @kanis_patel]

Inside the Agent Harnesses of Anthropic, OpenAI, LangChain, and Stripe https://t.co/nKt5gqETvW

[Kanishk Patel @kanis_patel]

@samzliu Nice! Do share if possible. Adding ASR would be great but achieving low latency would be tough in my opinion. But definitely a good addition.
I did add the ability to perform web research claims and make them stronger.

[Kanishk Patel @kanis_patel]

I often write articles to learn new things and share what I have learnt. With AI it has been easier but at the same time my learning from the articles have also reduced so I came up with Mati, it is a custom WYSIWYG with an AI twist.

It pushes back it asks me questions and tells me if my claims are unsubstantiated? It helps me avoid AI sycophancy and make sure I learn and share the right things. Do you think this is something that people will use.

Attaching a demo below.

Hoping for some feedback!

[Kanishk Patel @kanis_patel]

link to github repo -> https://t.co/Gdoly5I3UZ