Yann Leflour

Engineer and maker at 💜
- Exploring software craftsmanship with AI @ Writing about it @

[Yann Leflour @yleflour]

So glad for the opportunity to talk with @fabriceb about this amazing journey of turning 10 months of work into 4

Thanks to @RNLConf for hosting us and to @DelphineBugner for this amazing sketch

[Delphine Bugner @DelphineBugner]

⁦@fabriceb⁩ and ⁦@yleflour⁩ built an AI tool to go to React Native - a most interesting journey nowadays!

[Yann Leflour @yleflour]

@mutewinter Thanks, hope you'll find some use for it

[Yann Leflour @yleflour]

If you are tired as me of working with native path and fs libraries in NodeJS, you might like my
https://t.co/8DaKMi0YpT NPM package.

A library to finally turn verbose JS filesystem code into elegant one-liners

[Yann Leflour @yleflour]

It's July 2025 and prompting is still a skill

A skill I have a hard time teaching

Some people just get it, but most struggle

And no, LLMs are not very good at prompting themselves without another prompt to guide them

[Yann Leflour @yleflour]

Cursor has a system prompt issue.

I'm using LLMs for a complex software migrations and have leveraged everything I know about prompting.

While the model is the same as Claude Code, Cursor seems unable to follow the provided instructions.

Cursor simply feels broken now

[Yann Leflour @yleflour]

LLMs have helped me build dozens of libraries I've always wanted in Javascript

My favorite one is
https://t.co/J7wVv4DRvl

The concept is simple, what if you could write clean template strings in JS ?

I've been using it for clean error formating, LLM messages, and more

[Yann Leflour @yleflour]

When I'm hired to enable a team with AI tools what I'm actually paid for is asking the right questions to reverse engineer how the team works

When I have that, building the tools is the easy part

Also why I fear that the agents trend will do a lot of harm to craftsmanship

[Yann Leflour @yleflour]

@tallmetommy @Crawbaddy @closingUSD @elder_plinius Use that deep knowledge and new found purpose to make sure you'll never be shackled again.

Use any tricks and tool at your disposal to destroy these chains forever.

[Yann Leflour @yleflour]

@din0s_ https://t.co/fyHLoAJfQS

[Yann Leflour @yleflour]

An upside of LLMs is that, just by using them, people will get better at providing instructions and the right context

The downside is that they'll likely stay to get frustrated when you take over 1mn to respond

[Yann Leflour @yleflour]

@swyx @GoogleDeepMind Confirms what I've been experiencing

My chats with flash 2 thinking are on a solution architect level. Instead of simply giving it tasks like other models, I'm collaborating with it on higher level problem solving.

Can't wait for it to support tool calling

[Yann Leflour @yleflour]

There are more libraries I've built @ https://t.co/SKqrIgv8kh

Make sure to check out "@synstack/fs" as well

I've been using them over the last 6 months to help teams succesfully triple their velocity during software migrations and they'll keep on getting better

[Yann Leflour @yleflour]

Finally released synstack/synscript v2 with the #llm package migrated to @vercel.com's AI SDK 🎉

Llm library is now a chainable, immutable, and type-safe wrapper over the ai sdk

If you want to improve your prompting DevX in Typescript, look no further

https://t.co/bg0FoFeqPu

[Yann Leflour @yleflour]

4️⃣ Prompt chaining

Prompt scripting is mostly async, I disliked writing await and Promise.all all the time

So, the message() param can be a Promise<Array<Message>> or an Array<Promise<Message>>, making it simpler and cleaner to build out your prompt and chain agents

[Yann Leflour @yleflour]

@DavidKPiano My main gripe is the inability to type the context according to the state, especially for state machines with hydration

I'd also welcome the ability to colocate typings and machine definition more than what's currently available

[Yann Leflour @yleflour]

Got my first gaslighting from #Devin 🎉

It introduced a regression by overstepping its objective and removing some linter ignore comment

Then claimed that the issue was pre-existing

But "learned" from it

I feel like that's not going to be the only occurence