Alice Alexandra

Alice Alexandra

@alicealexandra.com

she/her • developer exerience, builder.io • writes, paints, composes, codes at alicealexandra.com • am generally, myself, a leaky abstraction

95 followers174 following135 posts

Running multiple AI agents gets ridiculous fast. - You have 100 windows open - Your computer fans are crying - Agents keep bonking into each other in the code - You can't ever shut the laptop, which makes it tough to just... have a life Here's how to manage the chaos: builder.io/blog/ai-age...

We keep saying the tech industry is losing its junior devs to AI, but there's a more nuanced take: Vibe coders are the next generation of juniors. In my technical writing these days, I'm thinking a lot about how I can teach and meet these newcomers where they are.

A helpful tip for AI coding: after a long task is complete, ask the agent how working in the codebase went. Then, turn any recurring feedback into issues for later agents to look into. This helps keep your codebase cleaner. Agents know what context they need that they're not getting.

In OpenCode, I've even made it a command: /survey: "How was working in the codebase today? Did anything trip you up with how things are organized? Are there refactors needed? Did your tooling let you do everything you needed today?"

Cursor's visual editor is cool for solo devs, but on a team, devs add zero value by moving a button 10px left. Designers should own those changes. Here's how the new design mode works, and what still needs to change before Cursor can be a Figma killer. www.builder.io/blog/cursor...

builder.ioCursor’s Design Mode (Visual Editing) ExplainedIs Cursor’s design mode a Figma killer? Here’s how it works, where it falls short for non-developers and on source mapping, and how to let designers edit code.

Anthropic agreed to pay $1.5B to settle claims it trained AI on millions of pirated books. But the judge hasn’t approved it yet, questioning if $3,000 per book is the right precedent. What do you think? Does this move us forward on fair compensation or just leave questions?

Personally, I think authors should be able to opt out of training on their works entirely. If their work is used, they should get options for ongoing compensation based on actual AI usage. (Though we don't have accurate ways to measure usage yet.)

It would be nice if you could prototype with AI right on your live site... ...without hand-replicating the repo, knowing how to code, or migrating your entire stack. And then send that PR right to your devs to approve. Luckily, you can. www.youtube.com/watch?v=8q3...

youtube.comProduction-grade AI PrototypingTry Fusion: https://www.builder.io/fusion?utm_source=youtube&utm_campaign=protoyping&utm_content=amSkip the mockups. Use AI prototyping to launch real, produ...

loving the DX of the new SvelteKit async updates. the best frameworks don't hype. they cut away ceremony, so things like async just... feel like you think it should be. fewer footguns, more flow. LLMs just "get it" reading the docs, too.

If you're shipping AI without evals, you're running prod on vibes. At Builder, we wire evals into CI, A/B tests, and prod to catch regressions and prove AI feature upgrades actually help our users. Here's how to set up your own. www.builder.io/blog/ai-evals

builder.ioYou need evals to ship AI featuresAI features are unpredictable and traditional tests fall short. Evals, automated checks for AI behavior, help you prevent regressions and measure success.

TDD feels bad. If I know how to write the function, why start with the spec? But AI agents switch it up. They need guardrails, and tests let them iterate on even the most complex problems. I'm convinced AI TDD is the way. Here's what changed my mind: www.builder.io/blog/test-d...

builder.ioTest-driven development with AILearn how AI transforms test-driven development (TDD) from a time-consuming chore into your secret weapon for building robust and bug-free applications.

Weird anti-AI scraping technique: digital tar pits. Websites hide links that lead bots to an endless maze of auto-generated gibberish that all link back to itself, poisoning training data. Wouldn't recommend, but still an interesting tactic against overly eager scrapers.

Figma's *official* Dev Mode MCP server is now in open beta, and it's a pretty awesome tool! I've updated this article with how best to use it, how its AI works under the hood, and where some alternate workflows can augment your design to code process. www.builder.io/blog/figma-...

builder.ioDesign to Code with the Figma MCP ServerTurn Figma designs into code using MCP servers. Skip screenshot guesswork and let AI access structured design data directly through Figma's API.

Debugging dark mode issues in an old, messy SvelteKit codebase is usually nightmare fuel. This morning, I connected my repo and put Builder Fusion on the case. It agentically iterated with the code and full DOM context, and it figured out the problem with a single prompt.

Alice Alexandra reposted

The word “login” comes from throwing a log attached to a rope with knots overboard a ship to see how many knots go by over time (see also, knots as speed). You’d then put that info in the “log book.” You’d “log in” on a regular basis. This wasn’t from 1959, it was likely from 1689! Etymology baby!

AI agents are like Twitter devs: great solo, terrible at collaboration. MCP gave agents standardized tool access to information across the internet, but the story isn't done. A generalist agent with tons of tools still isn't as useful as an orchestrated network of specialists.

Teach the model once, and then let it work. In other words: fine-tuning. - 50 great examples > 5k-token prompt. - Tone of voice and tool calls get baked right into the model weights. - Adapters make this possible for all. No second mortgage for GPUs. www.builder.io/blog/fine-t...

builder.ioFine-tune an LLM: Why, when, and howFine-tuning LLMs can save tokens, guarantee output formats, & bake in edge-case fixes. Learn when to fine-tune & how to do it effectively for your AI projects.