Hazel Weakly
@hazelweakly.me
I have thoughts. Lots of thoughts. They never stop thinking. Never stop thunking.
hachyderm.io/@hazelweakly
hachyderm.io/@hazelweakly
Pinned
Hazel Weakly
@hazelweakly.me
· 23d
The two hardest problems in Computer Science are
1. Human communication
2. Getting people in tech to believe that human communication is important
1. Human communication
2. Getting people in tech to believe that human communication is important
Reposted by Hazel Weakly
I showed Gas Town to my business partner Rich and he said “oh it's SOAP/WSDL for agents” and I am still processing that.
January 16, 2026 at 6:44 PM
I showed Gas Town to my business partner Rich and he said “oh it's SOAP/WSDL for agents” and I am still processing that.
This book has been one of my favourites to read and contribute to. I’m really looking forward to seeing the entire book in the flesh!
(And speaking of forwards… Be sure to read the forward in this book. I might be biased but I think it’ll be great 👀)
(And speaking of forwards… Be sure to read the forward in this book. I might be biased but I think it’ll be great 👀)
You'll note a new co-author on the cover, @aparker.io 👏
We have a stellar lineup of guest contributors, case studies, and deep dives.
The material is almost entirely new (maybe 5-8% carryover from 1st ed?), and at least 40-50% longer.
Books will be out by June, with preview chapters much sooner.
We have a stellar lineup of guest contributors, case studies, and deep dives.
The material is almost entirely new (maybe 5-8% carryover from 1st ed?), and at least 40-50% longer.
Books will be out by June, with preview chapters much sooner.
January 16, 2026 at 7:40 AM
This book has been one of my favourites to read and contribute to. I’m really looking forward to seeing the entire book in the flesh!
(And speaking of forwards… Be sure to read the forward in this book. I might be biased but I think it’ll be great 👀)
(And speaking of forwards… Be sure to read the forward in this book. I might be biased but I think it’ll be great 👀)
Reposted by Hazel Weakly
Just read this from @hazelweakly.me. Now, I can't wait to read the rest of Hazel's writing. If you're into #designsystems give it a look...and I'm guessing this line will resonate: "people who become known for seeing systems often become in charge of them."
hazelweakly.me/blog/to-be-a...
hazelweakly.me/blog/to-be-a...
To Be a Leader of Systems | Hazel Weakly
Picture with me, if you will, the absurdity of finding yourself swimming in the middle of the ocean. First think about the ocean and how deep and infinitely...
hazelweakly.me
January 15, 2026 at 7:19 PM
Just read this from @hazelweakly.me. Now, I can't wait to read the rest of Hazel's writing. If you're into #designsystems give it a look...and I'm guessing this line will resonate: "people who become known for seeing systems often become in charge of them."
hazelweakly.me/blog/to-be-a...
hazelweakly.me/blog/to-be-a...
GenAI tooling is weird. It’s “just” an intervention. Because it’s a tool. Any tool introduction is an intervention. Which means they work the same way any intervention work:
1. Understand the preconditions and culture norms required for success
2. Set people up for success
3. Respond to feedback
1. Understand the preconditions and culture norms required for success
2. Set people up for success
3. Respond to feedback
How do people feel about the usage of GenAI tools being mandated at work?
January 14, 2026 at 7:57 AM
GenAI tooling is weird. It’s “just” an intervention. Because it’s a tool. Any tool introduction is an intervention. Which means they work the same way any intervention work:
1. Understand the preconditions and culture norms required for success
2. Set people up for success
3. Respond to feedback
1. Understand the preconditions and culture norms required for success
2. Set people up for success
3. Respond to feedback
How come nylons only rip right before you’re going somewhere you need to look cute? Can they smell excitement or something? UGH
January 13, 2026 at 5:09 PM
How come nylons only rip right before you’re going somewhere you need to look cute? Can they smell excitement or something? UGH
Reposted by Hazel Weakly
One if the toughest things to judge when considering a new job or a new role is, basically, “Will this crush my soul?” Here’s a simple list of questions you can ask to help improve the odds that your next gig will be less soul-crushing: anildash.com/2026/01/12/w...
How to know if that job will crush your soul - Anil Dash
A blog about making culture. Since 1999.
anildash.com
January 12, 2026 at 10:25 PM
One if the toughest things to judge when considering a new job or a new role is, basically, “Will this crush my soul?” Here’s a simple list of questions you can ask to help improve the odds that your next gig will be less soul-crushing: anildash.com/2026/01/12/w...
Where’s the “they’ll join but muted and with camera off and they’re just there to talk massive shit about the office politics and drama on signal with their other work besties” label?
(Disclaimer: I am uhhh not in this picture. Yeah. Totally)
(Disclaimer: I am uhhh not in this picture. Yeah. Totally)
Google Meet show how late people will be based on meeting history
January 12, 2026 at 9:31 PM
Where’s the “they’ll join but muted and with camera off and they’re just there to talk massive shit about the office politics and drama on signal with their other work besties” label?
(Disclaimer: I am uhhh not in this picture. Yeah. Totally)
(Disclaimer: I am uhhh not in this picture. Yeah. Totally)
Reposted by Hazel Weakly
A single observability strategy often fails as companies grow. Discover how to build a meta-strategy that adapts from startup to enterprise.
By @hazelweakly.me, thanks to @embrace.io
By @hazelweakly.me, thanks to @embrace.io
Taking Your Observability Strategy to the Next Level
A single observability strategy often fails as companies grow. Discover how to build a meta-strategy that adapts from startup to enterprise.
bit.ly
January 11, 2026 at 3:30 PM
A single observability strategy often fails as companies grow. Discover how to build a meta-strategy that adapts from startup to enterprise.
By @hazelweakly.me, thanks to @embrace.io
By @hazelweakly.me, thanks to @embrace.io
Reposted by Hazel Weakly
Me when I just finished an entire book chapter about evaluating the quality of our evidence and designing to measure change and my tech social media is full of people angry at bad measures of change
a man in a hooded cloak is saying good your hate has made you powerful
Alt: Palpatine from Star Wars saying "good! your hate has made you powerful"
media.tenor.com
January 10, 2026 at 7:17 PM
Me when I just finished an entire book chapter about evaluating the quality of our evidence and designing to measure change and my tech social media is full of people angry at bad measures of change
It’s also, I think, one of the reasons AI is doing so well. There’s legitimate value to be had, sure, but when the vast majority of tech is one flavour of a scam or another, well…
It all but guarantees that AI is sold by scammers and to scammers. The utility of it is a happy accident, not an intent
It all but guarantees that AI is sold by scammers and to scammers. The utility of it is a happy accident, not an intent
This is exactly what I mean when I say that UX has no value when the business model is not "make a thing people want and sell it for more than it costs to make."
The top 10 most profitable tech business models these days are different kinds of scam. You don't need design for that.
The top 10 most profitable tech business models these days are different kinds of scam. You don't need design for that.
It's strange how far removed this Obama-era book feels from our time, with its maker-faire optimism during the heady days of the iPhone's first flourishing. But Don, I kept thinking, especially toward the book's flowery final chapters, don't you know everything is a scam now? Tech is scams now.
January 11, 2026 at 11:06 AM
It’s also, I think, one of the reasons AI is doing so well. There’s legitimate value to be had, sure, but when the vast majority of tech is one flavour of a scam or another, well…
It all but guarantees that AI is sold by scammers and to scammers. The utility of it is a happy accident, not an intent
It all but guarantees that AI is sold by scammers and to scammers. The utility of it is a happy accident, not an intent
Reposted by Hazel Weakly
This is exactly what I mean when I say that UX has no value when the business model is not "make a thing people want and sell it for more than it costs to make."
The top 10 most profitable tech business models these days are different kinds of scam. You don't need design for that.
The top 10 most profitable tech business models these days are different kinds of scam. You don't need design for that.
It's strange how far removed this Obama-era book feels from our time, with its maker-faire optimism during the heady days of the iPhone's first flourishing. But Don, I kept thinking, especially toward the book's flowery final chapters, don't you know everything is a scam now? Tech is scams now.
January 11, 2026 at 3:31 AM
This is exactly what I mean when I say that UX has no value when the business model is not "make a thing people want and sell it for more than it costs to make."
The top 10 most profitable tech business models these days are different kinds of scam. You don't need design for that.
The top 10 most profitable tech business models these days are different kinds of scam. You don't need design for that.
Reposted by Hazel Weakly
There's much talk these days about what's required for "real" intelligence, such as world models and the like. One thing we know about the human evolution of it: it required/requires emotion. That may have been better appreciated in 1938 than today.
archive.org/details/in.e...
archive.org/details/in.e...
January 11, 2026 at 7:51 AM
There's much talk these days about what's required for "real" intelligence, such as world models and the like. One thing we know about the human evolution of it: it required/requires emotion. That may have been better appreciated in 1938 than today.
archive.org/details/in.e...
archive.org/details/in.e...
Reposted by Hazel Weakly
When a public good is framed as too expensive it's stopped being seen as a public good. The university is so obviously a place and idea that should be owned by and for the public. 💔
Three short paragraphs, and you've got the whole mind-bending mess that is #UKHE finance & governance neatly laid out.
This is why it's all so exhausting: our managers declare there's only one static frame, while we know their framing is part of the issue.
💡 www.hepi.ac.uk/2026/01/10/w...
This is why it's all so exhausting: our managers declare there's only one static frame, while we know their framing is part of the issue.
💡 www.hepi.ac.uk/2026/01/10/w...
January 10, 2026 at 1:26 PM
When a public good is framed as too expensive it's stopped being seen as a public good. The university is so obviously a place and idea that should be owned by and for the public. 💔
Reposted by Hazel Weakly
My prediction is slightly different. I mean, it’s mostly this, but also there’s gonna be a very weird thing happening over the next few years:
AI is horrifically inefficient and too expensive. It’s just subsidised beyond belief. But that’ll end shortly (it kinda has to?). What happens after?
AI is horrifically inefficient and too expensive. It’s just subsidised beyond belief. But that’ll end shortly (it kinda has to?). What happens after?
My coding prediction for the year is that by end of 2026 agents will be good enough to write 90% or so of all code for production, and developers won’t all lose their jobs but there will be a prolonged mass grief event as they mourn the loss of hand-writing most code, an uneconomical activity
January 9, 2026 at 4:19 PM
My prediction is slightly different. I mean, it’s mostly this, but also there’s gonna be a very weird thing happening over the next few years:
AI is horrifically inefficient and too expensive. It’s just subsidised beyond belief. But that’ll end shortly (it kinda has to?). What happens after?
AI is horrifically inefficient and too expensive. It’s just subsidised beyond belief. But that’ll end shortly (it kinda has to?). What happens after?
Reposted by Hazel Weakly
I really do feel justified studying what exacerbates skill and role change identity threat and what protects against it wrt AI in software in 2023
My coding prediction for the year is that by end of 2026 agents will be good enough to write 90% or so of all code for production, and developers won’t all lose their jobs but there will be a prolonged mass grief event as they mourn the loss of hand-writing most code, an uneconomical activity
January 9, 2026 at 3:51 PM
I really do feel justified studying what exacerbates skill and role change identity threat and what protects against it wrt AI in software in 2023
Honestly we were always here, but you can cover up a multitude of sins by keeping humans in the margins just enough for a little plausible deniability. Once you remove that, though, the hard questions *must* be asked
They always should’ve been asked, mind you. But now they have to be
They always should’ve been asked, mind you. But now they have to be
Everyone's in the game of "have I actually motivated people to authentically care about my research and be real participants in it" now but some of us have been here for a long time eh
Have fun lol
Have fun lol
January 9, 2026 at 9:51 AM
Honestly we were always here, but you can cover up a multitude of sins by keeping humans in the margins just enough for a little plausible deniability. Once you remove that, though, the hard questions *must* be asked
They always should’ve been asked, mind you. But now they have to be
They always should’ve been asked, mind you. But now they have to be
Reposted by Hazel Weakly
I have never run a panel sourced, vendor survey project. I have always done careful and intentional recruitment from inside of communities. I defended this as a research leader when it was baffling to eg marketing and other functions. And meanwhile academics look down a lot of applied work. Well
“These findings provide clear evidence that data collected on MTurk simply cannot be trusted.”
January 8, 2026 at 10:02 PM
I have never run a panel sourced, vendor survey project. I have always done careful and intentional recruitment from inside of communities. I defended this as a research leader when it was baffling to eg marketing and other functions. And meanwhile academics look down a lot of applied work. Well
Reposted by Hazel Weakly
Unrelated ag fact: there are stingless Melipona bees that make honey!
Sometimes people think "I should raise those! Why do we even bother with honeybees, if we can just raise bees that don't sting?"
Then they try raising the stingless bees, & find out why almost nobody does that
Sometimes people think "I should raise those! Why do we even bother with honeybees, if we can just raise bees that don't sting?"
Then they try raising the stingless bees, & find out why almost nobody does that
They don't have capacity to terrorize more than 2-3 cities at a time.
Which means they rotate teams to new cities without relief.
Which means consistent resistance by fresh activists in each new city will break them over time.
Which means if you're city's not occupied, rest up and get ready.
Which means they rotate teams to new cities without relief.
Which means consistent resistance by fresh activists in each new city will break them over time.
Which means if you're city's not occupied, rest up and get ready.
"The Department of Homeland Security plans to pause operations in Chicago — where Gregory Bovino, a Border Patrol official, has led controversial arrest efforts — to support the immigration enforcement operation in Minnesota."
January 9, 2026 at 4:14 AM
Unrelated ag fact: there are stingless Melipona bees that make honey!
Sometimes people think "I should raise those! Why do we even bother with honeybees, if we can just raise bees that don't sting?"
Then they try raising the stingless bees, & find out why almost nobody does that
Sometimes people think "I should raise those! Why do we even bother with honeybees, if we can just raise bees that don't sting?"
Then they try raising the stingless bees, & find out why almost nobody does that
If I want to give talks on sociotechnical strategy, organisational design, or complex tradeoffs, what conferences might actually appreciate that?
January 8, 2026 at 3:10 PM
If I want to give talks on sociotechnical strategy, organisational design, or complex tradeoffs, what conferences might actually appreciate that?
Reposted by Hazel Weakly
New year, new newsletter! This one's all about the Liskov Substitution Principle and how it applies to more than just object oriented programming!
buttondown.com/hillelwayne/...
buttondown.com/hillelwayne/...
The Liskov Substitution Principle does more than you think
It's more than just the L in SOLID!
buttondown.com
January 6, 2026 at 5:21 PM
New year, new newsletter! This one's all about the Liskov Substitution Principle and how it applies to more than just object oriented programming!
buttondown.com/hillelwayne/...
buttondown.com/hillelwayne/...
Slept in my bed instead of on a random IKEA mattress on the floor for the first time in months. I missed this 😭
January 8, 2026 at 10:12 AM
Slept in my bed instead of on a random IKEA mattress on the floor for the first time in months. I missed this 😭
Oh look its a visualisation of me in a solo project that gets turned into a “group” project
This wheel is to the boy squirrel the way that the football is to Charlie Brown and it still makes me laugh every time he tries to use it only to be immediately spun.
January 7, 2026 at 1:01 PM
Oh look its a visualisation of me in a solo project that gets turned into a “group” project
Reposted by Hazel Weakly
Recommended read: off-the-shelf governance models for small FOSS projects.
antonin.delpeuch.eu/posts/off-th...
antonin.delpeuch.eu/posts/off-th...
Off-the-shelf governance models for small FOSS projects? | Antonin Delpeuch
antonin.delpeuch.eu
January 7, 2026 at 12:09 PM
Recommended read: off-the-shelf governance models for small FOSS projects.
antonin.delpeuch.eu/posts/off-th...
antonin.delpeuch.eu/posts/off-th...