Hello friends,

We have some incredibly exciting news, we’ve been nominated for an award! This community has touched thousands of people and changed lives (that’s not an exaggeration) so it’s wonderful to see that work recognised. Thank you to OpenUK for shortlisting us for their AI award!

This project started because Hannah selfishly wanted to be part of a community that was learning about AI together. There were existing tech communities but they were too deep into the engineering. There were other communities that focussed on users' needs - learn how to prompt this new thing and don’t worry about the magic black box. Then, there were of course the nonsense marketing events that made huge promises about what was possible without any substance. AI for the rest of us was built for the rest of us. Literally.

Don’t dumb it down, but don’t bamboozle me with jargon. Help me stay up to date. Help me stay grounded in reality. Help me have better conversations and make better decisions about AI.

Apparently Hannah wasn't the only person who wanted this community to exist! Thank you for being here!

If you’re based in London you can join the community in person on Tuesday 27th October. The theme of our next meetup is “The Agentic Web Explained” - A hot topic right now!

Have a wonderful week!

Hannah and Charles

Join Focused Focused are hiring both remote and office based roles in London, Denver and Chicago. Build agentic solutions with the best! View Roles

What’s Hannah reading this week?

A fraudster has just received an 18 month prison sentence for engineering a swarm of bots to stream his AI generated music, apparently receiving over $8m in royalties from popular streaming platforms.

Sentences on 6th October in North Carolina, U.S. Attorney Jamie McDonald commented:

This was clearly a direct attempt to make money by generating fake users, but it made me stop and think about some of the other ways creators make money on the internet. And how the mechanics of advertising, when it’s based on impressions and clicks, changes when so many of the users are AI agents.

Am I willing to pay per click when some of those clicks are agents? No, I am not. Or at least not as much. Do I care about impressions if most of those impressions are agents? I need to get through the agent to reach the user but how do I know when that’s happened?

I was reviewing the results of my LinkedIn experiment earlier this week, feeling quite proud that I’ve reached 15,574 users and generated over 38k impressions. Then I remembered my bot stalker (the account that comments on my posts with slop despite being called out for it) and started to wonder how many impressions are real human eyes and which are bots or automations. I wish I knew! I wish I could measure these impressions separately!

This week I was honoured to give the opening keynote at Fast Flow Conf in London. I’m always super critical of myself in these situations, never sure if I’ve done enough or landed the right messages. However, I think it’s a good sign when you get spontaneous applause half way through a sentence, and when every other speaker quotes your words back at you throughout the day.

I shared my thoughts on what I think about “Minimum Viable Human” (that’s a nerdy play on words for tech folks who are always talking about Minimum Viable Products btw). My own experiments and experience have led me to a place where I think WorkSlop and overuse of AI tools are potentially corrosive to our trust, communication, learning and decision making. I joke that so much of my work today relies on me showing up 100% human, authentic and original. I want to automate myself out of a job, I really do and yet I keep adding more work to the list of “stuff I do myself”. Minimum Viable Hannah is not so minimal after all.

My favourite talk of the day was Cat Morris and Nicolò Cambiaso Erizzo “Enter The Flow: Platform Product Management in the Era of Agentic Engineering” because they introduced me to the idea of “Intentional Friction”. It’s possible to automate every part of the software development lifecycle but Cat and Nico proposed that platforms should insist on human judgement and create just enough review and oversight to maintain quality.

This idea was also shared by Anthropic’s Engineering team in their post “Redesigning source control at Anthropic”​

If you haven’t read this story from Anthropic, go and do that immediately. Git became a bottleneck for Anthropic's engineering team and so they decided to build a replacement. Insane. Not only did they have to build a completely new source control system (called Scion) but they also had to migrate the whole organisation to it. A team of 7 engineers accomplished this unthinkable challenge in just 5 months. What planet are we even living on?

Cyber security has been headline news over the past 2 weeks. ASOS users received a suspicious message from hackers who addressed their global notification to the ASOS security team saying “Dear ASOS DPO and IT, we have fully compromised the Snowflake instance. Engage with us or we will leak it".

Then this week we’ve had another two high profile attacks. In France supermarket refrigerators were compromised, causing fresh produce to spoil, costing the supermarket chains millions and denying customers their usual groceries. Meanwhile Albert Heijm in the Netherlands suffered from a separate attack prompting them to disable many of their systems including parts of their supply chain and online ordering, leaving the Dutch government to consider emergency food reserves.

I’ll be donning my cyber-security hat this week at Kubernetes Community Days in Edinburgh where I’ll be talking about the Vulnpocalypse and how the threat of AI powered hackers is very very real and very very urgent. LeadDev’s latest article on the subject shares many of the same warnings and solutions. The answer is automation and the solution looks a lot like platform engineering.

Don’t forget, most cyber attacks don’t make the headlines. You’ll never hear about them. The increase in high profile, highly visible attacks is the tip of the iceberg, we won’t ever have a full picture.

The register reports on the UK government's response to the increasing cyber risks. It is still undecided whether the “Cyber Security and Resilience Bill” will be updated to address emerging AI risks. Many believe that it should be. Regardless of what the government says, the sensible thing would be to act anyway. Hope for the best and prepare for the worst, right?

In the world of software development Lenny’s Newsletter asked the question “How are tech workers feeling?” then Darren Fuller answered “Developers are not OK”. Both posts share the same sentiment, software teams are under pressure to deliver more than ever, because AI promised we could have more than ever. Burnout is surging, and optimism is fading.

In my talk at Fast Flow Conf this week I left attendees with a hopeful message. We already have the tools we need to navigate the changes we see happening all around us. If you haven’t reread the Agile Manifesto recently go and dust that down. It has stood the test of time. AI is making things fast, but is it speed at the expense of sustainability?

What's Charles reading this week?

In the UK, disability rights campaigner Baroness Campbell has become the first person in history to use an AI-generated voice in Parliament, during a debate in the House of Lords about the Government's plans for social care. Baroness Campbell has spinal muscular atrophy, which affects her speech. The Baroness has now created an AI voice clone that replicates the tone, rhythm, and pitch of her original voice using technology from ElevenLabs. In a blog post the firm says, “Our technology has now been used to restore the voices of more than 11,000 people who have lost their voices through injury or illness. We have committed to help one million people living with voice loss around the world.” Voice cloning technology is in many ways frightening: a friend of mine’s elderly mother was recently very nearly fooled by someone using a voice clone of him, and the scope for abuse is enormous so make sure you have a family code word etc. But, as this demonstrates, it can be a great accessibility tool.

Over in the US Character.ai, who we last wrote about in November 2025 when it was facing multiple lawsuits from families who say its chatbots contributed to teenagers’ deaths by suicide, has filed for Chapter 11 bankruptcy. After initially fighting the lawsuits, in January this year the startup agreed to settle several of them. In May this year, the Pennsylvania government also sued the company, alleging its chatbot falsely presented itself as a licensed medical professional in the state and gave a “fake Pennsylvania license number” to a user.

In 2023, Character.ai was valued at $1 billion after raising $150 million in a round led by Andreessen Horowitz (a16z). In a court document, Character.ai said it had both assets and liabilities of between $50 million and $100 million. The startup has agreed to sell some of its assets to Disney for $20 million, as well as a non-exclusive perpetual license for its technology. There’s not much detail as to what assets Disney wanted. (I have seen no mention of what happens to the data.)

In September 2025, the entertainment behemoth joined others in publicly criticising the firm for stealing content and harming children. At the time, Sara Fischer reported in Axios that Disney sent a strongly worded cease-and-desist letter that accused Character Technologies of weaponising Disney characters “in a way that could damage its brand long term,” and the scandal ended with Character Technologies wiping all Disney-owned characters from its platform.

Given this, Disney’s interest in the firm might seem surprising. However the House of Mouse has recently hired former Character.ai CEO Karandeep Anand, and is pushing into AI. Along with Anand, Disney told Todd Spangler at Variety that it expected “a number of Character.ai’s technical team” to join the company, prompting Robert Hart at The Verge to suggest that the move “resembles a kind of ‘acquihire’ strategy that has become increasingly common in Silicon Valley as a means of hoovering up startups’ talent that sometimes leaves the original company struggling to survive”.

Thinking of Disney, the $1.5 billion collaboration between it and game maker Epic (Fortnite, Unreal Engine), announced in early 2024, looks to be in trouble. Epic was slated to release a Marvel- and Star Wars-themed shooter in March 2027. That project has now been delayed indefinitely amid an AI-related shakeup at Epic and understaffed, slower-than-expected development thus far, according to a report in Bloomberg. Epic laid off 1000 staff in March.

Perhaps in response to Bloomberg's report, writing on ghastly X Epic CEO Tim Sweeney has said that:

Soon afterwards Epic announced a publicly available “Preview” build of its Epic Games Launcher, writing that “programming is hard. So, instead of doing that, we vibe coded a new Epic Games Launcher that’s way faster and more responsive than the current one, and has a new look.” There aren’t any details as to how Epic used vibe coding, and the cynic in me can't help thinking that coding a launcher shouldn’t be that hard for a software company. The vibe coded app “will have some bugs” that early testers can help squash with a feedback form, Epic admitted on the still ghastly X. The announcement post also warns that the vibe-coded Preview is currently “light on features”.

AI is clearly going to transform both the economics of creating games, especially the hugely-expensive AAA titles, and the mechanics of how games work since the NPCs (non-player or non-playable characters) can actually be intelligent now. AI also offers a new way for people to spin-up their own improvised games using tools like Google’s recently launched Playground. It will be interesting to see how it all plays out.

Updates

Early Bird Tickets Our next conference is set for 25-26 February 2027. Grab your early bird ticket for £150. Get Your Ticket

Between Jobs? Interested in a more human approach to career change and AI? Join the Surviving The Gap community for practical workshops, useful resources and early access to Career Core. Learn More

Follow us on LinkedIn Stay up to date on the latest news from the community! Follow

Follow us on BlueSky Stay up to date on the latest news from the community! Follow

Agent Craft Talks Agent Craft talks are now live! YouTube Channel