The talk I gave at NDC Oslo hosted by @NDC_Conferences became my biggest hit so far (10K+ views)!
Link: https://t.co/nEB1CP6EX9
I justify why advanced type safety constructs are more "worth it" than ever and show tricks for pushing agents into the "black hole of success" as @mattpocockuk would put it.
Give it a watch!
After 5 years and 100+ releases, Wasp v1.0 is in the works!
ETA: ~6 months
What's coming:
→ Every core feature reworked: Auth, RPC, Jobs, API, Deployment, etc
→ Full Stack Modules: install Stripe like a package, backend included
→ Native Windows, no WSL
→ Production-ready
Details (incl. roadmap): https://t.co/3UaoIwbrLT
My problem is that 98% of such contributions have obvious bugs.
They don't even bother to use AI that made the PR to review itself. That alone would catch most of them.
I'm fine with AI contributions, but at least also do an AI review first.
My hunch is that people don't bother spending their tokens on projects that are openly hostile or restrictive towards AI contributions. Nor according to where people who say things like "slop tokens" think they should go. https://t.co/leQZ6qnb6t
Another day of Unemployed Kyle (aka WDS) dunking on me because he has no real career experience.
His main complaint here is that I show multiple instances where I didn't know the depth of how part of a codebase worked.
Sounds bad until you realize it's a codebase with over 300 contributors and millions of lines of code. Doesn't matter how much AI is used, nobody knows the implementation details of all the code in a real world codebase like this.
The example he highlights is one where I was unfamiliar with details of how tagged skills are attached when we pass prompts to the Claude SDK. Funny enough, I know a lot (probably too much) about how we tag skills and how messages are passed over SDK, but this one intricacy was new to me.
Knowledge gaps like this are the norm in real world work. Learning to operate with partial understandings of systems is key to long term success as an engineer, regardless of AI usage.
I'm genuinely scared that inexperienced devs like Kyle are going to trick well-meaning juniors into holding back their careers.
Wasp is the first web framework to let you define an explicit "schema" for your web app.
Think of it as a database schema, but for auth, routes, cron jobs, data operations, email sending, and so on.
What your app does isn't anymore buried in the code, config files, and folder conventions, but has its own dedicated layer and lives next to your code.
@viborc I generally write my own prompts, and tune the "effort level" based on my needs.
I'd assume a pre-made prompts could have, since the most likely put some "scientific method guardrails".
It's hard to know what is good when vibe researching because agents are excited about every inquiry you make. Everything is a perfect in some use "case". You have to make negative statements to ground them constantly.
@dhh Most external AI PRs miss the whole point, so we have to reject them anyway.
These people also almost never use frontier models, nor do they check the PRs themselves. It just looks "okay" but is never actually good.
Let's see if Celld by @rough__sea can be "Cloudflare at home".
I've deployed @tan_stack Start on a local instance successfully:
• started Rustfs (S3 compatible storage)
• started Celld pointed at the storage
• built the app for Cloudflare Workers
And it worked... nice stuff! Slight modification to the Wrangler config and it deployed as-is.
What's the point you might ask, why not deploy the app directly at this point if you are going to self-host Celld anyways?
Yeah, you might want to host stuff your agents produced or others produced simply and quickly. Building a product around that primitive and such.