Romee Panchal

AI Architect | Writer, thinker, and lifelong learner. Subscribe to my newsletter:

[Romee Panchal @romitheguru]

Don't read to question it.

Don't read to agree with it.

Read to understand it.

[Romee Panchal @romitheguru]

Charging platform fee by e-comm by platform is like if a local shopkeeper ask customers a portion of rent for the shop

[Romee Panchal @romitheguru]

Coding agents are very good at building what you ask for.

The gap between what you specify and what you expect is where the agent starts making assumptions.

More assumptions → more rework → more iterations → more tokens.

The clearer the specification, the closer the output.

[Romee Panchal @romitheguru]

@nikks_techie It should always be tools/backend driven where agent only controls and updates the agent state

[Romee Panchal @romitheguru]

@nikks_techie I usually have state checked in the tool itself. For instance, when one tool is finished and it alters some state, and the agent decides which tool to use next based on the workflow, the next tool itself will check the state.

[Romee Panchal @romitheguru]

@maxintechnology @burkov Does the coding agent produce good quality Go code?

[Romee Panchal @romitheguru]

@Abhishekcur Fair points. I think the answer is to use AI more, experiment, notice what consistently works, and build processes around those patterns—just as we’ve always done with new technology.

[Romee Panchal @romitheguru]

Apparently, civic sense is not common sense.

[Romee Panchal @romitheguru]

Many people want AI to handle their thinking, execution, and everything in between—which surrenders their leverage. I think, AI should execute and retrieve information for me, but I keep the thinking.

[Romee Panchal @romitheguru]

An LLM is a next-token generation machine whose response is driven by system prompt + context + current query. If the requirements aren’t clear enough, you’ll just keep burning tokens without getting good results.

[Romee Panchal @romitheguru]

Many times users aren’t clear on their requirements for an LLM. They expect surprisingly good results even when they themselves don’t know what output they’re looking for.

[Romee Panchal @romitheguru]

@TZOpenCreators Exactly

[Romee Panchal @romitheguru]

Two agents can use the exact same LLM and still be radically different systems.

Give one:
full filesystem + network + shell

Give another:
sandbox + scoped files + restricted tools

Same intelligence. Different capability. Different blast radius.

Agent = Model + Harness

[Romee Panchal @romitheguru]

Is it just me, or does Codex hit its usage limits much faster than Claude for the exact same task under identical conditions?

[Romee Panchal @romitheguru]

The interesting part of “agentic video” isn’t that an AI can watch a 90-minute video.

It’s that it doesn’t have to.

Instead of processing every frame equally:

Question → search → candidate moments → inspect deeply → answer.

AI is starting to reason about computation.

[Romee Panchal @romitheguru]

Do you people change your coding workflow when a new SOTA model drops?

[Tapti Labs @TaptiLabs]

A useful rule in AI engineering:

Don’t start with the model. Start with the failure mode.

Ask:
• What can go wrong?
• Can rules solve part of it?
• Where does ML add value?
• How will we know it works?

Good AI starts with good problem framing.

[Romee Panchal @romitheguru]

@SwiggyCares Yes

[Romee Panchal @romitheguru]

Why do Swiggy's Bolt orders almost always take as long or longer than regular orders? I have never seen them deliver on time.

These people need to work on their estimation algorithms.
@Swiggy @SupportSwi88934

[Romee Panchal @romitheguru]

@SwiggyCares Here it is: #243694724166596

[Romee Panchal @romitheguru]

@SwiggyCares This is exactly the experience you are providing. Your customer support bot sucks. You don't fulfil your delivery promises and ask for a 100% cancellation fee.