A reminder, because part two lands Friday. Part one of my solo podcast series is the story of a 2016 pentest that came back full of CRITICAL findings, for HTTP headers I had genuinely never heard of. X-Frame-Options. HSTS. X-Content Type-Options.

More fun like this please tris.sherliker.net/blog/obfusca... Then again, randomly running base64 encoded shell scripts is very dangerous.

Something I have been thinking about, ahead of a post later this week. The first sentence someone says to you is roughly 14 percent of what they are actually thinking. The other 86 percent is still in their head, uncompressed.

It’s #CyberMentoringMonday!!!! Are you looking for a professional mentor or to learn more about InfoSec? Are you experienced and willing to ‘give back’? Use this thread and hashtag to connect

In case you missed it late last week: a new post on the shift from author to editor-in chief. It started when I heard a LinkedIn economist describe a model of work that matched something I had written in a case study months earlier. He has never worked in software. That is the interesting part.

Sunday thought. The best leaders diagnose what kind of conversation they are in before they respond. The most valuable developers know what to ask the AI for, and can tell when the answer is wrong. Both are skills. Neither is a personality you were born with.

Two things I wrote about this week, from two very different books, turned out to be the same idea. Duhigg: stories bypass the brain's instinct to look for reasons to be suspicious. Data activates that instinct. A good story slips past it.

New post today. I was listening to a Think Fast Talk Smart episode on careers and AI, and the guest described a model that mapped almost exactly onto something I had written in a case study months earlier. A thread on the shift from author to editor-in-chief.

Most technical change initiatives fail the same way. You gather the data, make the case, present it, and the team nods and changes nothing. Twenty years and fifty-plus organisations later, I see it constantly. A thread, from Supercommunicators.

In case you missed it last week: my first solo podcast episode since 2023. The story of a 2016 penetration test that came back covered in CRITICAL findings, for HTTP headers I had never heard of. It became an open source package now past 1.9 million downloads.

The most dangerous sentence in a design review is "I'm just thinking out loud." A thread on the authority gradient, and why senior people so often close down the conversations they believe they are opening.

Most of the problems that derail technology teams are not technical. They arrive dressed as technical problems. A thread on Charles Duhigg's Supercommunicators, and why the most useful thing a tech leader can learn is how a conversation actually works.

Sunday thought. The safest teams are led by people who go first, and say "I got that wrong" before anyone else can. The most expensive security gaps are the ones nobody taught you to look for. And purpose is still the question you cannot ask an AI to answer for you.

Yesterday's episode was about secure defaults. Here is the flip side: the day a default cost me a weekend. A thread on why "don't accept the defaults" is the most useful four words I know.

Apparently, I'm a nice person (which is what the grammaticaly incorrect "woke" means).

New episode, and a rare one: a solo monologue, no guest. My first since 2023. It is the story of how a failed penetration test in 2016 sent me down a rabbit hole that became an open source package. A thread.

S08E22 - From Pentest to 1.7 Million Downloads, Part 1: The Headers I'd Never Heard Of A pentest report full of CRITICAL findings — for headers Jamie had never heard of — sent him down a rabbit hole that ended at OWASP. In this first half of a two-part story, Jamie unpacks the HTTP security heade

In light of more terrible games layoff news, a periodic reminder that I do free headshots for job hunting friends for your LinkedIn/professional page/whatever. If you're in the greater Seattle area and need some new headshots let's chat

Most one-to-ones are a standup with one person. How is the ticket, any blockers, anything I should know. Twenty minutes, box ticked. That is not a one-to-one. A thread on the most underused tool a leader has.

I don't think I'd like to be known as a speaker. I'm more of an incoherent mumbler.

You cannot ask a team to be vulnerable in a place where vulnerability has never been modelled. The team is watching, before they decide whether it is safe to follow. A thread on going first.

Most founders build a WHY the wrong way round. They ask "what do we want to become?", pick an inspiring answer, and reverse engineer a purpose to fit. A short thread on why Simon Sinek says that almost never holds.

I did a thing today. Feels good to be helping out the next generation of computer scientists get hands on experience with technologies.

Careful folks. The last time this happened, we left the EU.

Most companies will tell you they have a blameless culture. Most of them do not. A thread on what one actually looks like, built around a first day of consulting I have never quite shaken off.

And there's a new Chillhop seasonal compilation?! Heck yes! chillhop.bandcamp.com/album/chillh...

You'd best believe that I backed SKYWALKER: A Fan-Made Rap Tribute by @megaran.com on @kickstarter.com www.kickstarter.com/projects/meg... I just dislike that it took me this long to back it.

Sunday thought from this week. Purpose is the question AI cannot answer for you. The sceptic on your team is usually the colleague who, when they turn, brings the rest with them. Small, deliberate wins out-deliver the flagship most weeks.

A Wednesday lunch meeting turned "that's not possible" into "we shipped it last week". That is the short version of a Community of Practice that worked. A thread on the format that moves knowledge between teams the org chart was designed to keep apart.

@codrina.bsky.social's tip for getting started with tiny on-device AI models: the Hugging Face catalogue, plus the Microsoft Foundry playground. Whisper for speech. LLaVA for image descriptions. Custom Vision will give you a bespoke detector from about fifty training images.

I try not to use hyperbole when I share links to videos, and I'm certain that I'll lose friends by saying this, but my word! This video by @techconnectify.bsky.social is fucking awesome. Its timely, its fantastic, and chilling all at the same time. www.youtube.com/watch?v=KtQ9...

The engineering team did not show up to the AI training. Operations did. Finance did. Sales ops did. Six months later, the engineering team was watching their non-developer colleagues outpace them on productive intuition. That was the lever.

As someone who is having to deal with the very real issues with a neighbour having planted bamboo, I can tell you that no one should ever plant bamboo. It was planted a year ago, and it's already 2 miles from their house and chocking the foundations of one of the other houses on the street.

I know it's from a time that most of us would rather forget, but I'd like to start using Liz Truss as a scientific method for measuring how long something lasts. Just like the lettuce that outlasted her time as Prime Minister.

Most best-practice advice from an AI assistant is the average of the public internet. Average is often fine. Average is not always best. A thread on the part of the WHY that cannot be outsourced to convergence.

A consultant cannot convince a sceptical senior engineer in a town hall. They can sometimes convince them in a pull request. A thread on how an AI sceptic at Ligentia ended up daily-driving the tooling he had argued against.

When AI can generate a working codebase in an afternoon, the WHAT is no longer the hard problem. Knowing what to build, and why it matters, is the part that stays expensive. A thread on Simon Sinek's Start With Why, for engineers.

Catching up on some episodes of @loadingreadyrun.com's CheckPoint from last month, and I have to say that I appreciated the throw-back joke about the Wii Proctology t-shirt. I still have mine somewhere.

Sunday thought from this week's episode with Codrina Merigo. ( @codrina.bsky.social ) The interesting question is rarely "can AI do this?". It is "does this need the frontier model, or would a tiny one on the device do the job, more privately and for less?

Ten years; lots of no’s; endless networking; huge doubts; countless red eyes in middle seats; and 100k words later: the smile of a woman who just sent the final pages of the completed manuscript of her first traditionally published book to her editor!🥹 “Leading with Empathy For Dummies” drops 9/29!

Architecture is the set of decisions you wish you could change later but cannot, cheaply. A thread from my CPD review of Fundamentals of Software Architecture, for the teams making those decisions right now.

My question to train companies in the UK is this: what's the point of producing a timetable for your services? It seems like the trains areove (or not) based entirely on vibes? I know that everyone is doing their best, and there are things beyond their control, but I feel like they could do more.

New episode of The Modern .NET Show. Codrina Merigo (codrina.bsky.social) joins me to talk AI-enabled .NET MAUI apps, and to challenge the assumption that "adding AI" means bolting on yet another chatbot.

S08E21 - Beyond the Chatbot: AI-Enabled .NET MAUI Apps with Codrina Merigo Developer, international speaker, and book author Codrina Merigo joins Jamie to discuss her new Apress title ‘AI-Enabled Apps with .NET MAUI’. Together they look beyond the chatbot — covering hybrid online/offline architec

The job was never the typing. Donella Meadows' Thinking in Systems is the CPD book I most often hand to engineers moving into leadership. A thread on why the system, not the code, is usually the thing to debug.

Your quietest standup might be your most expensive one. When a development team stops surfacing problems, the problems do not stop. They just stop being visible until they are costly. A thread on psychological safety.

The dream of the 6̵0̵s̵ ̵7̵0̵s̵ ̵8̵0̵s̵ ̵9̵0̵s̵ ̵2̵0̵0̵0̵s̵ ̵2̵0̵1̵0̵s̵ 2020s

Hey, it's the podcast I work on. Thanks for the boost @dot.net!

I’ve struggled with panic attacks, depression, and agoraphobia for decades. I have published over 20 novels with different publishers and made friends around the world. This is just to say: you can still do great things with mental illness. I believe in you. ♥️