iiinova — where your digital self keeps evolving

Reposted by Michael McCliment

Dr Abeba Birhane @abeba.blacksky.app · 07/10/2026

you don’t hold "AI" accountable; you hold AI developers, vendors, and corporations that profit off it accountable holding "AI accountable" is like holding a toaster accountable. it's an inanimate object. it doesn't have wishes, feelings, or desires

301658514

Reposted by Michael McCliment

Molly White @molly.wiki · 06/10/2026

“While OpenAI admits to agents behaving ‘unpredictably’, they must also acknowledge their responsibility to monitor and prevent these risks,” Wikimedia said. “AI companies are not doing enough to secure their systems and protect the public from the harm they cause.” arstechnica.com/security/202...

arstechnica.com

OpenAI agents tried to hack Wikipedia tools and flooded it with traffic

The reports of OpenAI agents harming third-party sites keep coming.

1330194

Reposted by Michael McCliment

Eric Maugendre @data.yt · 17/05/2026

“The most familiar example of integrating as the social process is when two or three people meet to decide on some course of action, and separate with a purpose, a will, which was not possessed by anyone when he came to the meeting but is the result of the interweaving of all.” Mary P. Follett, 1924

122

Reposted by Michael McCliment

Dan McQuillan @danmcquillan.bsky.social · 30/09/2026

"Anthropic leaders to control AI lab via 'Founder LLC' to promote the common good" www.reuters.com/legal/transa.... If only Elinor Ostrom was still around to explain to these AI doomer numpties that 'common good' is something we decide _in common_, not via a cabal.

screenshot from linked Reuters piece which reads: "Sept 28 (Reuters) - Anthropic has long positioned itself as the virtuous AI company. An early look at its IPO filing shows how it is creating a new "Founder LLC" vehicle aimed at serving the common good, while insulating its leaders from market forces.
The new LLC will be made up initially of Anthropic's seven co-founders including CEO Dario ‌Amodei, individuals who are "distinctly equipped to be stewards of our mission," according to a copy of the filing, seen by Reuters. Anthropic's mission is to benefit humanity through responsible AI."

02310

Reposted by Michael McCliment

Kelsey Hightower @kelseyhightower.com · 28/09/2026

I hope we keep writing docs for people who want to understand and build things without using AI.

1431953

Michael McCliment @cornazano.bsky.social · 20/09/2026

First time in Germany! Unfortunately not getting out of the airport, though.

Picture of me at the back of a very long line to get through border control.

020

Michael McCliment @cornazano.bsky.social · 19/09/2026

Second trip to Malmo for the year. I learned today that YYZ has an infield concourse, you have to take a shuttle bus to get to the real gate.

000

Reposted by Michael McCliment

Ben Kuchera @benkuchera.bsky.social · 16/09/2026

When I was writing about the four day work week I went into a deep dive on the research and psychology behind optimal work and everything I learned completely changed my life. Resting is literally the most efficient way to get good work done, and it's not even close.

29027

Reposted by Michael McCliment

Mitch Dyer @mitchyd.bsky.social · 15/09/2026

“Crunch is inevitable” guys are so funny, just openly admitting they can’t and won’t make a production plan

241900426

Reposted by Michael McCliment

Anthony Moser @anthonymoser.com · 10/09/2026

this is, I think, one of the inherent problems with LLMs using natural language interfaces: they are a statistical tool *that looks like* a logical tool statistical approaches can be a useful way to find related information! BUT for it to be useful you have to *know that's what it's doing*

938293

Reposted by Michael McCliment

headfallsoff @headfallsoff.com · 08/09/2026

the next generation of game developers don’t need a special grant for 20 geniuses to get some walking around money they need entry level jobs that a) exist and b) keep existing next quarter

61215277

Reposted by Michael McCliment

tef @tef.bsky.social · 06/09/2026

update 2: this point has come up a few times and i think it's worth reiterating here code review, albeit adopted as a cheap way to find bugs, is more about knowledge transfer in practice. doing the review at the end of the work was just the easiest way to introduce it into a process

2295

Michael McCliment @cornazano.bsky.social · 05/09/2026

I finished reading Argyris and Schon’s Organizational Learning II during this week’s work trip, so I (finally) got to start on Cat Hicks’ book on the flight home!

A copy of The Psychology of Software Teams sitting on an airplane tray table, about to be opened for reading.

000

Reposted by Michael McCliment

Mar Hicks @histoftech.bsky.social · 03/09/2026

Anyway, remember that things are inevitable right up til the moment they’re not. “Inevitable” describes potential, not reality. It is a nightmare of the future coming to bully those who are asleep—not a description of some kind of logical or pragmatic reality, as some would have us believe.

219347

Michael McCliment @cornazano.bsky.social · 30/08/2026

Heading to Montreal for the week. Flying this time, which means a bit more time with my family on the weekends before and after.

Jetway and some Air Canada planes at YYZ.

110

Reposted by Michael McCliment

Chanda Prescod-Weinstein 🌌 @chanda.blacksky.app · 27/08/2026

Here’s the thing I will say about books, in their paper format: they don’t have ads or pop ups in the middle. There are no notifications that will pop up on the page. It’s just you and those sentences and those ideas. You’ve got to think. There’s no likes or comment section. It slows you down. Good.

39217

Reposted by Michael McCliment

Moon Blight 🫀 @cuntmoney.bsky.social · 01/07/2024

Words are empty, unstable signifiers that have clouds of meaning and connotation. People don’t like to think that we put parts of ourselves into those meanings too, but we do.

0153

Reposted by Michael McCliment

Olivia Guest · Ολίβια Γκεστ @olivia.science · 09/08/2026

Chess is a formal system. Language isn't. There's a huge difference between the rules of chess and language, which doesn't have a set of rules in the same way.

3628

Reposted by Michael McCliment

Olivia Guest · Ολίβια Γκεστ @olivia.science · 08/08/2026

exactly 💯 on LLMs: they simulate the output we call language, not the cognitive capacity we call language!

492513702

Reposted by Michael McCliment

Dare Obasanjo @carnage4life.bsky.social · 17/07/2026

There’s a big difference between the promise of GenAI and the reality. The promise is the democratization of expertise. Anyone can be a coder, analyst or lawyer with AI. The reality is actually the erosion of expertise. Being an expert is no longer valued yet experts are needed to vet AI lies.

1511420

Michael McCliment @cornazano.bsky.social · 17/07/2026

Starting a three-week vacation. I'm pretty sure this is the longest proper break I've taken in over 25 years.

030

Reposted by Michael McCliment

Patricia Aas 🐢🏳️‍🌈 @patigallardo.bsky.social · 09/07/2026

This is one of the things I «discovered»* about programming, that you can have an «accent» from your more proficient languages in your less proficient languages. Getting Good at a programming language is much more than learning the syntax, it’s cultural * ie was never taught and hadn’t read about.

9317

Reposted by Michael McCliment

Anthony Moser @anthonymoser.com · 03/07/2026

LLMs are a kind of uncertainty filter that give credible sources and bad ones the same glossy finish, severing the links so you can't really be sure

526073

Reposted by Michael McCliment

Anthony Moser @anthonymoser.com · 04/07/2026

I'll add that knowing who said something and why they might have said it is a basic and extremely important kind of context which means LLMs are functionally *designed to destroy context*

620852

Reposted by Michael McCliment

Robin-Yann Storm @rystorm.com · 30/06/2026

The conclusion I pull from this isn't that AI is great at its job, it's that the bureaucracy and politics of corporate environments are so terrible that folks will try to find any possible way out of it. Because AI can mean no more recalcitrant bob engineers that you have to deal with. No politics.

181

Reposted by Michael McCliment

transmutrix @transmutrix.bsky.social · 29/06/2026

remember when a bunch of software companies made a big deal about "returning to the office," citing worsening quality of work? and then none of their products got any better? and then some of them started to vibe code everything? and their products got even worse??

241024194

Reposted by Michael McCliment

Mike Bithell @mikebithell.bsky.social · 27/06/2026

This is a lost art. First place I worked had 4 teams working on different games, and a dedicated pitch team just laying track in the form of new projects. Used to be doable at 200 people. Indies being so small and AAA going all in on big projects has largely nuked it. It’s a shame.

932835

Reposted by Michael McCliment

Patricia Aas 🐢🏳️‍🌈 @patigallardo.bsky.social · 20/06/2026

Typing speed was never our limiting factor. We are distancing us even more from our users, from the collaborative process of creating what is needed.

1342

Reposted by Michael McCliment

Cat Hicks @grimalkina.bsky.social · 16/06/2026

Doing this can be more than a bit terrifying because it requires you to have some collective intellectual humility; in other words, a culture of admitting out loud and in public that you don’t know everything, and that you need other people’s help.

291

Reposted by Michael McCliment

Dr Abeba Birhane @abeba.blacksky.app · 14/06/2026

I understand and sympathise with the lack of funding to study the impact of ai on kids/vulnerable groups. regardless, if you're partnering with/sponsored by the ai company whose technology you're studying, the conflict of interest is too great to take your findings seriously

210217

Michael McCliment @cornazano.bsky.social · 10/06/2026

Managed to read a chapter of nonfiction during my commute today. First time this year, which is a decent sign that I’m getting to a healthier point with work/life. When my personal load-shedding behavior involves not reading, it’s a good indicator that I’m teetering too close to burnout.

010

Reposted by Michael McCliment

Sarah Drasner @sarahedo.bsky.social · 18/05/2026

I’ve decided that my anxiety is just my brain consistently doing premortems

41279

Michael McCliment @cornazano.bsky.social · 18/05/2026

In Malmo for the week, beautiful weather and getting to meet colleagues in person for the first time at the Massive studio. 😊

The outside of the Massive office in Malmo, with the name on a rainbow series of flags.

020

Michael McCliment @cornazano.bsky.social · 17/05/2026

About to get on my first flight since 2019, and it’s also going to be the longest one I’ve ever taken. First time heading to Europe. 😃

030

Michael McCliment @cornazano.bsky.social · 06/05/2026

New Reading for during the commute to work this morning. 😃

A physical print copy of Platform Decay, by Martha Wells.

000

Reposted by Michael McCliment

Camille Fournier @skamille.themanagerswrath.com · 02/05/2026

This is for real also I'm hiring SRE-shaped people if you are hands-on SRE not "thinks SRE is mostly about telling people what to do"

76218

Reposted by Michael McCliment

ve, the ghost of info past @vortexegg.com · 21/11/2025

Shannon’s entire thing was borrowing metaphors of human activity to use as nouns to stand in for unrelated physical processes and then not differentiating between the two.

141

Reposted by Michael McCliment

Asawin Suebsaeng @swin24.bsky.social · 20/11/2025

I know I’m a libtard but I think it’s bad the U.S. president keeps calling for the killing of specific sitting elected officials who didn’t even commit crimes

390160312744

Reposted by Michael McCliment

Kirill Osenkov @kirillosenkov.com · 15/11/2025

AI PRs where the author doesn't have an understanding of the system shift the burden of understanding onto the reviewer. So now it's on the reviewer to gain an understanding of this area to either confirm or refute. The author saved time at the expense of the reviewer.

111729

Reposted by Michael McCliment

Nick Tune @nick-tune.me · 05/11/2025

I never saw enough potential at that point for recovery. It's hard to recover when bad choices are baked in and trust has been severely damaged. If you're building an internal developer platform, you need to work closely with teams from the start.

021

Reposted by Michael McCliment

Yvonne Lam @yvonnezlam.bsky.social · 11/10/2025

highly recommend @ruthmalan.bsky.social's Papers in Systems. We haven't updated the website in a minute, but our monthly virtual meetups have restarted after the summer break: papersin.systems/about

papersin.systems

About

A friendly, virtual journal club

272

Reposted by Michael McCliment

Trond Hjorteland @trond.hjorteland.com · 07/10/2025

Suspect many will get a hands-on experience of what the ironies of automation is all about with all this LLM assisted coding.

052

Reposted by Michael McCliment

Wiredferret! @wiredferret.bsky.social · 02/10/2025

Missing data is a fucking tragedy. It’s so easy to write a bad survey and get bad data back.

121

Michael McCliment @cornazano.bsky.social · 30/09/2025

My current reading dips back towards the cognitive science part of my interests. I did not expect to get ambushed by Taylorism in this context, but here I am.

Two paragraphs from The Extended Mind discussing the effects of neo-Taylorism on cognitive performance.

020

Reposted by Michael McCliment

Cat Hicks @grimalkina.bsky.social · 23/09/2025

- a question I like to ask is "what are the common ways this goes wrong" or "what are the common misconceptions people have here." Good teachers become experts in the *misconceptions* not just the "right answers". See if you can identify just one for everything you're writing about!

1278