Cary Thomson

Founder of Agent Ash - An AI coding agent for Mac. @agentashdotcom

[Cary Thomson @cary_thomson]

Introducing Agent Ash - An AI Coding agent for Mac.
https://t.co/qfOspA3u3R

[Cary Thomson @cary_thomson]

@Wattenberger @migueldeicaza I’d call it a “meadowmap”.

[Cary Thomson @cary_thomson]

@perrymetzger The best thing is that it actually *answers* the question that you asked. Search returns results *about* the search terms you entered. That saves me a lot of time.

[Cary Thomson @cary_thomson]

@mitchellh How much has LLM usage extended the scope/complexity of the tasks that you tackle?

[Cary Thomson @cary_thomson]

@DavidSHolz “Touch grass”. Trite but true.

[Cary Thomson @cary_thomson]

@geoffreylitt I like this framing!

[Cary Thomson @cary_thomson]

@stevesi Metaphors are powerful and useful. The problem is if we start believing LLM’s are more than tools.

[Cary Thomson @cary_thomson]

@jarredsumner @theo Really? Can’t you feel @theo’s pain just thinking about it.

[Cary Thomson @cary_thomson]

I iterate. I’m around 4-5 on the first pass(es) because I value getting working code early. Then down to 2-3 for a review/tidy pass(es). What do you do?

[Matt Pocock @mattpocockuk]

The "should you read code" debate is dumb because the real decision isn't binary, it's a scale:

1. Reading every line of every diff
2. Scanning every diff, reviewing important lines
3. Ignoring diffs but understanding the 'why' of every PR
4. Spot checking PR's instead of reading every one
5. Ignoring PR's, but doing regular spot checks on the codebase
6. Ignoring the code, but spot checking agent traces to help improve the system
7. Ignoring both the code and the system, let models handle everything

Where are you on the scale?

[Cary Thomson @cary_thomson]

@sawyerhood “Argument from authority”. Works on people too.

[Cary Thomson @cary_thomson]

@deanwball @AndrewCurran_ I agree with this. Of course no-one really knows where and how fast model capabilities will grow, but we have no reason we are there yet and the sky’s (not even) the limit. Exciting times.

[Cary Thomson @cary_thomson]

@GergelyOrosz I consciously do this, sometimes the metaphors hit a sweet spot. (Seams is a good example).

[Cary Thomson @cary_thomson]

@charliermarsh Is this your way of telling us you won the lottery, without out telling us you won the lottery ? 🤔

[Cary Thomson @cary_thomson]

@int32max Be nice to your AI, you don’t want to be first against the wall when the revolution comes.

[agent-ash @agentashdotcom]

Did you know - Ash has its own built-in Xterm compatible terminal, complete with retro '80s vibes.

You can also !one-shot bash commands, but sometimes you need to drop into a full terminal session.

[Cary Thomson @cary_thomson]

Working on some visualisations . I find @observablehq awesome but fiddly. The AI is crushing it. I can be much more ambitious and achieve more - faster.

[Cary Thomson @cary_thomson]

@mitsuhiko I know what you mean. I found a bug the other day because my tests were suddenly slow. Turned out my wifi connection was shitty. Never would have found the bug otherwise.

[Cary Thomson @cary_thomson]

@perrymetzger Agreed, also I don’t want to be first against the wall when the revolution comes. 😜

[Cary Thomson @cary_thomson]

@mitchellh My next recommendation is Winnie the Pooh. That bear has good taste, in honey.

[Cary Thomson @cary_thomson]

Some really good advice in here. Recommended reading.

[Angus Norton @NortonAngus]

AI in Product Management: The Prototype Is the Easy Part https://t.co/9SuYIlsvSu #productmanagement #ai #aiprototypes