finer world

AshLua has been released! Easily and safely provide #AshFramework actions to a lua script to be called in your application 😎. Works via the very excellent lua library by @davydog187! See the guides for a very special element: integration with AshAI. Easily provide MCP tools for reading the docs of your new custom lua SDK as well as evaluating a script 🤯

Kenya cannot tax itself into prosperity. It must produce, export, invest, manufacture and empower enterprise.

A core tenet of #AshFramework is to "Model your domain, derive the rest". And the "rest" is often some kind of code generation or compilation step, which makes your actions accessible over some interface. AshGraphql builds an Absinthe schema for GraphQL support. AshJsonApi makes your actions available over Phoenix controllers or plugs. AshTypescript emits a typescript client. Ash has had introspection tools since the beginning. You can say things like `Info.attributes(Post)` to get a list of attributes. However, we ended up repeating the same complex patterns in every extension of Ash to take some set of actions that should be exposed, and answer questions like "what are all the types we need to expose?" or "what filters/sorts does this resource support?" Thanks to Torkild (Ash core team member), we now have a tool that makes this *extremely easy*. It's called `%Manifest{}`. This manifest can take a given set of actions, i.e "entry points" (or just "everything in an app") and gives you all of this information right at your fingertips. Each item in this data structure has a place for you to put extension-specific information "right where you need it" for later usage. Over time we will back-port this new data structure into our existing extensions, but moving forward this is by far the best way to do code generation with Ash!

Learn to negotiate. Rent, suppliers, credit terms, delivery costs and payment timelines can make or break you.

Business is not just buying and selling. It is discipline, timing, relationships, data, trust and cash control.

The future belongs to businesses that understand data, customer behavior, costs and digital visibility.

Before you had AI to spit out large swaths of code, you had to write it all yourself which gave you a natural in-depth understanding. With AI, the path of least resistance now gives you an easy way out of understanding the code you are shipping. But the fact that you shouldn't ship code you don't understand *hasn't changed*. You just have to resist the temptation 🙅 On the flip side of this, AI provides all kinds of opportunities for understanding that were never reasonable before. Whenever I'm having trouble grokking a large code change, I use this simple prompt and so far it has yet to fail me: > Take each non-test hunk from the git diff, and create an HTML page slide show for each one, explaining why it had to change and where it fits in the big picture. Put it in an order that will help me understand this change. In less than a minute I have an understanding aid that would have been entirely infeasible in the past. The only flaw here is that you have to be really honest with yourself when reviewing AI generated code. Do I understand this well enough to speak to it later? Am I confident in the second order effects of this code? Have I thought through the edge cases? If not, don't just "ship it anyway". Use your fancy new tools to help you understand it 😎

How to ACTUALLY get your first 100 users This guy grew his saas to $3M/year with a playbook anyone can copy: 1) Ranked against every competitor and stole their traffic 2) Built free tools. Embedded his product inside 3) Built demos for strangers on Reddit for free 4) Showed up everywhere his users were Now? 150,000 users. And counting.

~1 month ago we filmed a new project for @starter_story with @marclou we weren't 100% sure how it would turn out... new ideas and formats are always so hard to predict. But I'm SUPER proud of the outcome: the story and the style are awesome. (shoutout @editorartem srsly, best in the game) can't wait to bring you more stuff like this thanks for trusting us with this project, Marc 🤝 Go give it a watch!

Just added discord to loopr too 😎. Spent ~6 years of my life building lego bricks with #AshFramework and damn it feels good to actually build something with it. Wildly validating feeling. With Ash and LLMs this stuff is just comically easy.

At this point, this is just irresponsible. Yes, coding agents are leading to an increase of software production, but we are not seeing a similar push or increase in software quality. If Anthropic focuses on safety and it believes software engineering is going away, then it needs to be doing much more to improve how we design, build, test, and maintain software (aka software engineering). Increasing the production of unreliable, poorly designed, and unverified software directly undermines safety. Claude Code is claimed to be "fully written by AI". In the last two months, it took three separate postmortem-worthy failures and user complaints to surface what their own testing missed. Yesterday users were being over billed by hundreds of dollars. Software engineering isn't ready to go away and there is not enough progress to argue that case. I am certain Anthropic would argue that AI progress in other domains is strongly dependent on having proper safeguards in place. I can't wrap my head around the cognitive dissonance when it comes to software. PS: Mythos (may) improve software security, but that is only a subset of safety.

A simple website with a good domain name might just be all you need.

@dashboardlim Claude design

I've created a full guide on how to design anything with Claude Design You also get exact copy-paste prompts for interactive immersive websites, animated pitch decks, app mockups, social banners, infographics, and pricing pages Grab it FREE Like + Comment "CLAUDE DESIGN" and I'll DM you the full guide No opt-in, no BS

This is such a fantastic video explaining recursive language models(RLM) in a very succinct and clarifying way. It also shows why using it with @DSPyOSS makes it such an amazing choice. Great job @isaacbmiller1 youtube.com/watch?v=AsgDgY…

The workflow I use to ship production software with AI coding agents. Not 10 agents in parallel. Just 2. One implementing a feature, one prepping the next. Coordinated through Tidewave's task board. I built Elixir Radar's newsletter system in a week this way. It's live, real customers are using it. Full walkthrough:

jumping on this right away!

To help developers get started with the Claude Elixir Phoenix plugin, I've added the phx:intro command. It explains the basics, how to use the plugin, and what's under the hood. Here's the first section. #Elixir #PhoenixFramework #LiveView 🧵 1/6

Someone built a web-based System Design Simulator, where you drag & drop architecture components and actually simulate traffic, failures, latency, and scaling in real time, System design just got way more interactive.

Very suspicious that OpenAI and Anthropic both dropped major announcements within hours of me publishing this: dashbit.co/blog/why-elixi… They want to silence it. Don't let them. Read it twice. Share it with your colleagues and friends.

Supervision trees are fault tolerant. Now they are self-aware. Introducing Beamlens: Adaptive runtime intelligence for the BEAM. It embeds AI agents directly into your supervision tree to diagnose incidents, analyze state, and self-heal in real-time. Here is how it works 🧵👇