Renan

Building AI

[Renan @renancmw]

startup idea: jev but 1 million context window and zero data retention

[Renan @renancmw]

@alexdanilowicz same here! being able to feed customer insights back into the product and roadmap looks like a superpower

any resources you’d recommend?

[Renan @renancmw]

@shishirmehrotra do you have a router between different models? would love to know more about it

[Renan @renancmw]

Getting out of the model's way is the moat. Models are smart enough to figure out what they should do and gather the necessary context with a minimal toolset.

One of the beauties of pi is that it is really simple to extend, too. That’s true for AI products overall: be simple, nail your core, and let your users extend it as they need.

[Pi @pidotdev]

Over the past few months Pi has consistently shown up in benchmarks as the best harness for both cost and performance.

Earendil engineer @cristinaponcela examines case studies from @databricks and @Shopify highlighting why Pi’s minimalism is the advantage.

Blog post below

[Renan @renancmw]

I'm doing customer interviews this week, and I read The Mom Test by @robfitz in one sitting yesterday. If you're building a company, it's one of the most useful things you can read on how to actually talk to customers

[Renan @renancmw]

vibe-coding an agent is basically free. running an effective agent at scale is still hard.

- small component failures compound
- agents don't crash; they return weird answers
- security, permissions, and logs
- tool reliability and maintenance
- evals and regression testing
- control latency and cost
- retries, fallbacks, etc.

a lot of moving components

[Renan @renancmw]

@berprecht Agree, I'd add: exclusive partnerships, scale economics, network effect

[Renan @renancmw]

I'm really bullish on SML. Fine-tuning a really effective model will be the key to AI products

[Thinking Machines @thinkymachines]

Today, we are releasing Inkling-Small.

Inkling-Small achieves comparable performance to Inkling at a quarter of its size. It features 276B total parameters, 12B active. We are making the full weights available.

https://t.co/BtYNcpkDRA

Fine-tune it on Tinker today, or chat with it in text, image, and audio on Tinker Playground.

[Renan @renancmw]

@TakoTreba that's great! can you share the model you are using?

[Renan @renancmw]

I’m preparing for a startup launch, and it’s been a really interesting journey. Building the product is only 50% of the work.

[Renan @renancmw]

6+ months after the launch, I still haven't seen any MCP apps in prod. Is it worth it?

[Renan @renancmw]

reading crossing the chasm is one of the best playbooks for bringing an AI product to market

[Renan @renancmw]

@joeschmidtiv Nice post! For the lighthouse, I'd add: choose a niche small enough to dominate, then expand to adjacent ones that share the same problem

[Renan @renancmw]

@askOkara launch

[wesley⚡️ @wesleymagness]

coming out of this conference - i have to give it to a few companies who really outshined the rest:

turbopuffer - @Sirupsen
modal - @modal
planetscale - @PlanetScale
factory - @FactoryAI
exa - @ExaAILabs
openrouter - @OpenRouter

obvious ones (Anthropic / OpenAI / Cursor)

[OASIS @oasisdevices]

Today we introduce OASIS 1. 

The smart ring built for private dictation. Whisper to write. Touch to edit. 

A first step beyond the keyboard toward a world where your intent follows you across every device.

Order at https://t.co/gZieZw6vYJ first batch is limited.