when we go together, we grow faster!
Constant learning is my core.
Love driving AI adoption and vibe coding solutions that actually fix something at scale.
Read this section from Fable's system card.
We now have to worry about frontier models faking productivity while sabotaging work under the hood.
When a model has a tool setup, it will use that tool, even if with a low chance, Anthropic are poisoning prompts now.
@mattdotroberts The assembly of agents can be easily turned into an agentic workflow through any of the tools you mentioned, the important part is starting with thinking about what they will build
@mattdotroberts Today I was teaching our 9 year old, asked her if you want to build a game, what team do you need? She started with a planner and a coder, then after 5 minutes of back-and-forth she got to an assembly of a planner > Plan verifier > Coder > tester > fixer > enhancement planner
@trq212@bcherny@AnthropicAI Please remove or rename the "Workflow" tool!! this is such an annoying trigger after 4.8 release, don't you have a human review step when rolling these out? If so, that step seems to be flawed.
Most Al work does not happen in the tool
It happens before you touch the keyboard
The thought:
"What am I actually trying to create?"
The prompt is not the work
The thinking is the work
Don't underestimate thinking time. It is the one advantage no platform can take from you
@mattpocockuk Check Cerevisor, I built this as a tool for me to edit and refine these longer workflows.
Skills are an amazing tool, but we are always tempted to add more to them to get out of them.
Html has been good way for human comm ever since Gemini started being reliable, this says 2 things:
1- Anthropic late to the game on using html for human comms
2- They aren't relying on .md files for agent2agent handovers or they're pushing less technical folks to consume more
HTML is the new markdown.
I've stopped writing markdown files for almost everything and switched to using Claude Code to generate HTML for me. This is why.
@mattpocockuk I don't get the point behind AI detectors, a good writer with AI powers can create super valuable content with the help of AI, why would we penalize them for adopting the technology? And why would we pretend that something is not written by AI?
Working on a dictionary of AI Coding
Perfect for when you want to sound smart to your teammates/bosses
Here are the entries on the smart zone / dumb zone, and attention mechanisms: