Nirmal

Follows and tweets on AI - AI Transformation- Product Management* Public Cloud* Cyber Security*

[Nirmal @cloudynorm]

There is no better time to get into Cloud☁️☁️ related roles. If you are willing to learn and be curious opportunities are aplenty.Some free resources to get started. 🧵

[Nirmal @cloudynorm]

@CTOAdvisor A blog post please

[Nirmal @cloudynorm]

@rohanpaul_ai @chamath Haha he didn’t know that Anthropic is building the refrigerator and everything that’s gets inside

[Nirmal @cloudynorm]

@wulfie_bain_ @OpenAI Will you be open for Melbourne based candidates?

[tobi lutke @tobi]

https://t.co/5MJ7u1VHwf

[Eugene Yan @eugeneyan]

some thoughts on working with ai models

• context as infra
• taste as config
• verification for autonomy
• scaling via delegation
• closing the loop

https://t.co/pdd8bk66Jz

[Nirmal @cloudynorm]

@AbhiumNaanum Why

[Viv @Vtrivedy10]

The Harness is a Context Manager on Behalf of the Model

What happens when the context window fills up and who decides? This decision is external to the model - The Harness designer must have some opinion here! Decisions like this are crucial in turning a model into a great product for end users.

The context window is a sacred boundary beyond which all model computation actually happens. Context engineering is important because designing what gets passed over this boundary is the main determinant of agent performance. Harness design is how you decide how this boundary gets managed.

Our create_agent primitive in LangChain exposes the one of the simplest Harnesses for builders to extend, a ReAct loop with support for tools, middleware (hooks), and model choice

It’s a great place to start in agent building because it forces you to think through and contend with all of the design details that transform a simple agent loop into a purpose-built agent for your tasks

The first time you hit the context boundary in a simple agent loop, the API will just error out and your agent run will end. The API contract only supports a max number of tokens.

A harness helps you get in line with the API contract by managing context via strategies like truncation, compaction, offloading, and targeted context eviction

This is just one decision to think about in Harness Design, many more come up as you build such as agent specialization via Subagents, Tool design, Skill design, and more. Each of these are important in extending a model to make it into a useful product for users.

create_agent is a great level of abstraction to start building agents. Builders can go up a level to deepagents for a more out of the box agent experience and even further to Fleet as a more out of the box product experience. Or they can go down to the runtime execution level to LangGraph like @caspar_br had talked about

starting from a simple harness to build a great agent helps you learn fundamentals of how models work + good design patterns that turn them into great agents and products

[Nirmal @cloudynorm]

@mattsgarman @OpenAI Welcome Matt to X

[Jaya Gupta @JayaGup10]

https://t.co/yiGW8jKKXk

[Nirmal @cloudynorm]

@chamath @chamath - when are you expanding in Australia?

[Nirmal @cloudynorm]

@pristinaoffl Agents for US visa? Does this really work

[Nirmal @cloudynorm]

@meenakshei It was Claude for me all the time and feeling 5x productive in all things work and personal

[Nirmal @cloudynorm]

@CTOAdvisor Full stack AI steering specialist

[Nirmal @cloudynorm]

@introvertsmemes Yes

[Nirmal @cloudynorm]

An interactive, real-time visualization of how LLM context windows fill up — including MCP server overhead, KV cache state, context rot, and token distribution across message roles.

https://t.co/UDBzXb8c6C

[Nirmal @cloudynorm]

@ThoughtfulTechy Yes

[Nirmal @cloudynorm]

@realneerajnegi Message me

[Nirmal @cloudynorm]

@kris_sg @itihaasa I work in Bigtech ( FAANG) and it’s changed how we work , the biggest hit will be for IT services not for the people who do the actual work ( engineers etc) if your work is around managing the actual work there is no need - it reduces effort by 60-70% and many cases 5x-10x

[Nirmal @cloudynorm]

@theBuoyantMan Exactly I know someone in Big IT and never heard of Claude

[Nirmal @cloudynorm]

@amorriscode @mndoci