levi

also building another sickass tool at @lightreachbuild . any guesses on what this is?

@dudufolio progressive brrrrr

dark mode - progressive blurrrrrr

linear’s latest update makes one thing clear: the future of product development is less about tracking work, and more about moving it forward. that’s exactly why we’re building foundry from @lightreachbuild. from requirements.md to decomposed tasks, contextual assignment, and execution. if you’re thinking about the same problem, follow along. we’re building foundry in public and will be sharing more very soon!

ran brand refresh experiments for a client today if you’re curious about this workflow, drop a like + comment and I’ll share more

this might be one of figma’s most important updates in a long time can't wait to try this out!

@figma feels like a bigger shift is happening tho iin the future, are we still designing in figma, or designing through agents?

Now you can use AI agents to design directly on the Figma canvas, with our new use_figma MCP tool and skills to teach them. Open beta starts today.

@jakubkrehel it's always in the small details that define what “looks fine” and “feels engineered”. thanks for the tip!

One tip from my make-interfaces-feel-better skill is using `font-variant-numeric: tabular-nums` for values that update. It makes digits equal width, similar to monospace fonts. Keep in mind that some fonts change the appearance of numerals when this is used.

@kyleanthony clean af

Brand design for Tenkara

@felixleezd we could just be one skill.md file away from making joyful designs 🤠 kidding

Design needs to be joyful, not just functional. Bring back joy in design.

what used to take 2–3 weeks in figma now takes a day with claude code + codex the real unlock is the .md files if you want to 10x your design workflow like this, follow me. i’m sharing everything

@uixamjad looks really clean!

These are just beautiful masking and effects, you can simply create it in figma 🥳🥳🥳🥳

kinda bittersweet but its time to say goodbye to figma [for now] built this in two days just with claude code and codex for our leads pipeline if u want a tutorial on how i made this, reply in the thread!

@perkmaybe let's connect!

You can gain 400 followers per day. If you want to gain followers, drop "Hi" and connect with everyone who likes it. I’ll do shoutouts for my active followers.

@lunarfq hello!

NO account should be under 1k !! say hello and we follow you

s/o to my cracked devs @jcmerllin @lr_oludayo and a stealth ai/ml/math demigod glued to his desk 👽 follow @lightreachbuild for more updates

also building another sickass tool at @lightreachbuild . any guesses on what this is?

@delveroin building compress.lightreach.io frontier models at a fraction of the cost. we route requests to the most efficient model capable of doing the job.

Happy Thursday Founders! Share what you're building 👇

@mfranz_on @mfranz_on we also noticed a more important key to this puzzle working on compress.lightreach.io ....Issue is that the LLM IDEs run significant retry/harness logic in the background that our tool exposes to you

@CiprianiRanieri We are lowering costs and increasing observability with cursor and claude code usage by living at the middleware/router layer. Showing users exactly what prompts and retry logic are being sent to the LLM: compress.lightreach.io

Running a frontier model for every request is like using a supercomputer to open a PDF. Most AI tasks are simple. Classification. Formatting. Small transformations. Yet they’re processed by the most expensive model available. That’s where most AI cost comes from.

The future AI stack probably looks like this: user request ↓ router ↓ best model for the task Cheap models when possible. Frontier models when necessary. Efficiency becomes the default. compress.lightreach.io