jamie

A or B

I suspect it’s almost certainly because of chat/social network controls and age verification in some jurisdictions and their collaboration features could be classed as such.

Seems like a wider trend we’ll see as government go weird

'Hot desking' seems a far too sexy phrase for what is essentially just moving around the person who isn't worthy of office furniture.

Enjoying the For You feed? Give it a like ♡ to help more people discover it: bsky.app/profile/did:...

The more people use it -> the more feedback we get -> the better we can make it for you.

First example I've seen in the wild github.com/tldraw/tldra...

100%
LLMsare having the strange effect of splitting apart crafts which used to necessarily coexist. You can solve problems and build things without writing code (though programmatic thinking and understanding the code remain valuable).

I’m somewhat hopeful that people will find motivation to do hard things. I think a lot about how people go to the gym, or start running, learn to swim and the next thing is they’re setting PBs, doing their first 5km, marathon, or triathlon. Not easy things—hard things but self-motivated

I whipped up another learning related Skill! Smaller than Learning Opportunities, but very complementary: interactive guidance through a quick research-backed psychological intervention that helps improve learning plans, motivation and commitment



github.com/DrCatHicks/l...

This sounds like a joke but I’ve been on a flight from Atlanta that was delayed for 2 hours because the plane had the wrong charger cable for the pilot’s iPad (used for charts etc)

Will strip for half price electricity!

This doesn’t appear to be related to the DoD standoff (about how to models are used, but this was about how they are trained). I’m still hoping Anthropic hold firm on the use of their models but I’m less hopeful now

look they are a large corporation so I'm sure they're not sainted innocents and I'm sorry if I sound like a shill, but there clearly are in fact consequential differences between Anthropic and the other AI companies politically and morally and I'm tired of pretending that that isn't the case.

Tried out sprites last week and it was cool but within 24hrs the sprite was unrecoverable (according to their support)

There’s just this brief moment of arbitrage when some coding has become easier/cheaper and the market has not adapted to paying less for it.

I can’t see that lasting.

I’m kinda enjoying some agentic coding for the sheer “I can’t believe this thing which was only a dream is now a functional reality”.

And yet, I also can’t let go of the idea that valuable activities are, by their nature, hard and easy things aren’t

Yes, saw some lovely earth last week!

Still only love for Cork, right?

We had one in the garden this week!

It’s always a joy to edit some “proper” photos on the now-rare occasions that I get the DSLR out

Introducing Checkset

Checkset is a #ruby gem for repeatable verifications using Playwright.

Picture the most critical, common smoke tests you do after deploys or on PRs. Automate that and increase your confidence in your changes.

See the post for more information.

afomera.dev/posts/2026-0...

Entering the world of custom internal bots and, yep, that's a terrifying message

a group of people with the words is this the real life is this just fantasy on the bottom ALT: a group of people with the words is this the real life is this just fantasy on the bottom

Oh no

I may have jinxed myself since the sprite.dev I built the bot on seems to be broken and won't load.

So I rebuilt it on a better foundation in ~1 hour.

What do LLMs see?

I wrote a lil' tool that extracts the attention matrices out of open models and creates this typing visual, with each token's opacity changing according to its average attention score as the prompt progresses. Dimmer words are considered less important to the model.

That’s a thing I could only wistfully imagine a year ago but the effort to get there was always too great. Now I can basically invent things by accident.

And I haven’t even committed it to GitHub because if it died today I’d have it rebuilt in an hour.

It’s a wild, weird, destabilising feeling

Yesterday in about 4 hours, including being distracted by meetings and lunch, I built a bot that can respond to a Slack thread with an AppSignal error, investigate it in AppSignal, pull in production logs, and understand our codebase, then produce an initial theory. It can then create a Linear issue

Frantic is a good word. Perhaps a bit manic too.

I think this is what happens when the gap between imagination and implementation is basically zero—and all the tumbling consequences of that for our industry.

Book cover of "Designing Data-Intensive Applications, 2nd edition". It has a similar wild boar on the cover as the first edition, but it uses O'Reilly's new cover design, and the boar is now slightly colourised.

The second edition of Designing Data-Intensive Applications, by myself and @chris.blue, is finished and sent off to the printers! Ebooks should be available in the next week, and print books in 3–4 weeks. Sigh of relief. 😅

(BTW, this is a good opportunity to support your favourite local bookshop!)

Last night I dreamt that Cam McEvoy’s
(the Australian Olympic swimmer)

dog
(I don’t know that he owns one)

was chasing and attacking sheep
(whose sheep? I don’t have sheep)

so I needed to tell him that the dog needed to be put down
(why me?)

Anyway, I guess I might be stressed

Hard same. The business aspect is the thing I’m most worried about—and not for today, but in the months and years ahead. Like, it’s still sunny now but it looks like rain on the horizon and I’m not sure what the best shelter looks like