Paul Johnston

Paul Johnston

@pdj.bsky.social

Multi CTO, Climate, Data, Governance, Government, NED, advisor, strategist, ex AWS and GDS Other: Green Party/ADHD/Leadership/Christianity M: @pauldjohnston@mastodon.green X: @pauldjohnston

927 followers365 following5.2K posts

Pinned

'In other words, we basically prove the adage "YOU CAN'T BRING FACTS TO A FEELINGS FIGHT"' Love it

Kate Klonick@klonick.bsky.social

Bigger Finding # 2: Outrage-evoking misinfo seems to have little concern for truth It is shared more w/out reading, but people still could tell you it wasn't true, they just don't care In other words, we basically prove the adage "YOU CAN'T BRING FACTS TO A FEELINGS FIGHT" 7/

Paul Johnston reposted

good post and roundup by @jkriggins.bsky.social - what's next for DORA and the tech industry in measuring engineering team performance? i see Atlassian DX and AWS as key centers of gravity. Will be closely watching what @lauratacho.com and @nathenharvey.bsky.social cook up at AWS.

leaddev.comWhat’s next for DORA?Google has stopped running the DORA survey, leaving engineering leaders without an independent benchmark as AI makes productivity harder to measure.

Paul Johnston reposted

ppl screaming in my replies that AI does nothing while I can only move my fingers to type this post because of how good I was at using AI tbh

Rothaí@rothaiocht.bsky.social

Wait til the markets find out Anthropic’s valuation is based on Claude giving me magic words to make my health insurance pay my claims.

I get this. I really do. There are LOADS more games out there now, and clearly a lot of games created with AI. Skills never made businesses. Money does.

Bruno Dies@brunodias.dev

please stop replying to game devs saying things like "well look on the bright side those laid off devs can start their own indie studios now" because eventually one of us is going to snap and "WITH WHAT FUCKING MONEY" would be a shitty last thing to hear before you die

Paul Johnston reposted

Replying to @grimalkina.bsky.social

The other thing I wish we could admit is that every side of this could in a given case be wrong. Doctors can be wrong. Science can be wrong. Patients can be wrong. AI can be wrong. Whether we get to the best outcome with the information at hand has to be emergent

Amy Feldman@amyfeldman7.bsky.social

*NEW* Colossal Biosciences’ Ben Lamm and Anduril’s Palmer Luckey team up on a new startup to genetically engineer war dogs for the military. Warhound comes out of stealth at $1.2B valuation For @forbes.com with @iainmartin.bsky.social and @mhunterhart.bsky.social www.forbes.com/sites/amyfel...

The US is now a terrible terrible place

Amy Feldman@amyfeldman7.bsky.social

*NEW* Colossal Biosciences’ Ben Lamm and Anduril’s Palmer Luckey team up on a new startup to genetically engineer war dogs for the military. Warhound comes out of stealth at $1.2B valuation For @forbes.com with @iainmartin.bsky.social and @mhunterhart.bsky.social www.forbes.com/sites/amyfel...

MAKE IT A POSTER AND SELL IT

HulloThere@jethroreading32.bsky.social

Three years ago, my dad said that all fish are the same shape. Out of spite, I decided to draw all 478 families of bony fish listed in Nelson 2016. There's still some stuff I need to edit and some fish I want to redraw but, after 985 days, I finished my big dumb fish project last night.

I am testing out posting again on X. A lot of the people I would expect to leave have not. And if anyone tells me that I'm in some way supporting Elon Musk then I will respectfully ignore your opinion.

Paul Johnston reposted

Right then, so what are you actually stockpiling? Years ago, owing to a run of generous visitors from various points east, I had about 20 litres of finest vodka on hand which I figured could see me through for a while...

The rule now is: Don't use open source unless someone is being paid to maintain it. If not, be prepared to fix every security hole.

How it left Russia

RSPB@rspb.bsky.social

Goldeneyes make a splash in today’s Autumn Arrival Spotlight. 🎩 As they arrive from Scandinavia and Russia, look for their striking golden eyes and crisp black-and-white plumage on winter lakes.

Paul Johnston reposted

So who / what's buying the free lunch? via @aaron.bsky.team

Dave Willner@dwillner.bsky.social

When every category costs another model call, teams make uncomfortable trade-offs: lumping distinct harms into one blunt classifier, checking only a sample, or quietly leaving some rules unenforced. With CoPE-B-S1 those trade-offs more or less disappear.

Paul Johnston reposted

Huh. Everything is slow. Not least Claude (at least for me). Maybe everyone is trying to patch everything and nobody wants to do it without Claude.

The idea that this is "slashing red tape" is funny. This is the main reason for us not getting an electric car up to now. Victorian terraces need this, and that's a LOT of UK housing stock.

Margot Hodson@margothodson.bsky.social

Good news for those wanting an EV without off-street parking: www.independent.co.uk/news/uk/home...

I love how everyone thinks that you just need to ISO the crap out of everything and you'll be fine. Are you good at cybersecurity? *Hands over ISOxxxxx* Jeff over there is an idiot though.

One of the things I've learned in recent months, is how you simply cannot rely on a supplier to disclose that they have been attacked. Their business is not your business. Their stakeholders are not your stakeholders. Assume every supplier will only disclose when forced.

Andrew Couts@couts.bsky.social

NEW: AWS detected attempts to exploit the PHP vulnerabilities on October 17, four days before they were patched. Logs indicate multiple malicious actors tried to use the flaws. @lhn.bsky.social reports: www.wired.com/story/hacker...

Paul Johnston reposted

I do think it's kind of funny to be someone who's been thinking and posting about like, human productivity and problem solving and the workplace for so many years before AI because people just cannot believe you could independently be interested in it forgive the millennial GIF but

I'm up to 5 *BIG* cybersecurity stories this week. And by big I mean just one of them would have the potential to be on mainstream news in a major country. And none of them are directly because of any of the AI firms which is amazing.

Paul Johnston reposted

The Govt claiming we must drill more fossil fuels so we can use the tax revenues to fund the green transition is about as logical as a doctor telling you to smoke 2 packs a day to raise the money for your lung transplant You don’t put out a fire by pouring on more fuel 🔥#Jackdaw

Paul Johnston reposted

the fact that an intermediate CA in India that sells only into the Indian market, when compromised, can generate TLS certs for literally any website on the internet, that feels like a major trust boundary problem

Filippo Valsorda@filippo.abyssdomain.expert

After the firefighting is done (and a nap) we'll need to have a long, serious conversation about the state of non-browser non-platform X.509 verifiers. The Linux status quo is "a pile of PEM files behind a dumpster" and that's not an acceptable excuse for a root store in 2026.

Paul Johnston reposted

Not Laura Kuenssberg promoting the idea that local councillors are helplessly watching BBC 1 on a Sunday morning for guidance rather than... actually just democratically elected local leaders very capable of making their own decisions.

I'm sick and tired of Green Party members telling me that "Zionism is Racism". It's so unbelievably toxic as a statement and the way that many justify it sounds very much like antisemitism. It's naivety in the extreme. But they've been emboldened by the motion passing.

Seriously wondering if I can just build a startup, sell it, and check out of tech. It wasn't much fun before and it's even less now.

Paul Johnston reposted

It seems a lot of people have no idea what's going on in regard to open weights The oligarchs stand against open and free technology. They want your rent. There's a significant movement amongst whom "alignment" and "trad values" are synonymous. They've played the anti-AI crowd for absolute fools.

In other news, I may be looking for something new in the next month or so. Not totally sure yet. The VC might want me to build a template for future startups for them.

Paul Johnston@pdj.bsky.social

I've been working hard for a VC for the last 3 months or so building out a new startup. I'm handing it over to the new team they've brought in (their decision) and it's fascinating to see how easy and difficult it is to hand over an AI assisted solution. When the AI is seset up, it's fine.

I've been working hard for a VC for the last 3 months or so building out a new startup. I'm handing it over to the new team they've brought in (their decision) and it's fascinating to see how easy and difficult it is to hand over an AI assisted solution. When the AI is seset up, it's fine.

View full thread

One thing that really matters is underlying principles. It's something many tech people misunderstand. I use AWS serverless as much as possible. If you come from a kubernetes background it can be very confusing in understanding the philosophy of building a serverless solution. AI doesn't help.

If you ask "why not kubernetes?" it will say "because the principles say but to". Which means you have to give underlying reasons for your principles. And that's hard. But it's also the most valuable part of the decision repo. Basically I'm handing over my philosophy of development with AI.

I've been working hard for a VC for the last 3 months or so building out a new startup. I'm handing it over to the new team they've brought in (their decision) and it's fascinating to see how easy and difficult it is to hand over an AI assisted solution. When the AI is seset up, it's fine.

View full thread

The other aspect is that I've built this on serverless principles and the team coming in is not as experienced in that as I would like. (I get that I'm one of the experts and wouldn't expect my level of knowledge but still ...) So handing over is a challenge.

One valuable aspect I've built is a decision repository. Separate from all code repositories, it's simply a way of explaining why the solution is built the way it is. Point Claude Code at it and you can easily find out why the solution has developed from the foundations.

I've been working hard for a VC for the last 3 months or so building out a new startup. I'm handing it over to the new team they've brought in (their decision) and it's fascinating to see how easy and difficult it is to hand over an AI assisted solution. When the AI is seset up, it's fine.

Getting the new team to understand why and how the AI is set up the way it is, is challenging. There is very little standardisation in the way that AI is used. Some people think that means agent swarms and harnesses. Others think it means code generation and not much more.

The point is that it's entertainment. If it was a game designed to be "winnable" in a "if you figure out the tricks and patterns" way, then certain people would win all the time. But it's not designed to be won that way. It's designed to be played to avoid being banished or murdered.

Sam Freedman@samfr.bsky.social

This is the first series I've watched and it is weird that such a badly designed game can be so compelling.

View full thread

Late game: No idea who the traitors are. Just pick one person you are sure is a faithful (as much as you can). Fifth rule: get to the final round table and don't worry about catching traitors Sixth rule: team up with the most likely faithful left. Remove everyone else. Hope you're right.

The point is that it's entertainment. If it was a game designed to be "winnable" in a "if you figure out the tricks and patterns" way, then certain people would win all the time. But it's not designed to be won that way. It's designed to be played to avoid being banished or murdered.

Sam Freedman@samfr.bsky.social

This is the first series I've watched and it is weird that such a badly designed game can be so compelling.

The game is therefore about group psychology. First rule: Don't stand out in the early game. Second rule: build multiple alliances and hold them loosely. Third rule: follow the crowd when voting unless it's pointing at you, then acknowledge it and have two targets ready.

Mid game changes a little. Alliances are more set. Traitor murders are more targeted generally but they try to make it random. You never catch traitors unless they make mistakes. Fourth rule: accept you'll get it wrong and focus on one person at a time to limit your exposure to murder.

Paul Johnston reposted

I find this fascinating. Zack has showed that social media can gain a huge following and build a party. Similar to what Farage has shown in reality. That popularity allows you to build a movement. The question is whether it's the right movement, or whether he has lost the philosophical roots

Ian Dunt@iandunt.bsky.social

Holborn and St Pancras byelection: Polanski is an utter failure and it's time for him to go iandunt.substack.com/p/holborn-an...