Join us on Wednesday 7th October as AI Exchange takes over AWS Farringdon for a night that’s set to be one of our biggest yet.

**[AWS](https://aws.amazon.com/?nc2=h_home&refid=ea4f1bda-c329-47b3-a6d0-278045ca3aea)** are bringing along ARI, their humanoid robot to the AI Exchange for a live look at just how far physical AI has come. From robots breaking human records to the everyday tasks they still can’t master, this talk dives into humanoid robotics, World Models and what it will really take to close the gap between humans and machines.

We also have a fascinating talk from **[Faculty](https://faculty.ai/en-gb)**, sharing how they’re using AI to tackle terrorist propaganda online, helping platforms automatically detect Daesh content and deploy the technology securely.

**[Aries Global](https://www.aries.global/)** are diving into why good documentation and coding principles aren’t always enough when agents take over, and how harsh linters, static code checks and architecture enforcement can make sure AI generated code actually meets your standards before it gets anywhere near production.

Doors open at 6:30pm and talks start at 7:00pm sharp. RSVP as soon as you can to secure your space.

\*Please note - There is a potential space for a third speaker too. If anyone’s keen to get involved and has something exciting to share, please [click here](forms.gle/gMBEyAyyLmA1iQAe7) to submit a talk.

**Agenda:**

**[Ben Atkinson](https://www.linkedin.com/in/bgmatkinson/), GenAI & Agentic AI Specialist @ AWS**

**Humans vs. Humanoids: Who Wins — and Do World Models Change Everything? Featuring ARI the humanoid Robot**

A robot just broke Usain Bolt's 100m record — then crashed because it couldn't stop.

Is that the story of physical AI in 2026? superhuman at the spectacular, hopeless at the everyday. A robot can sprint faster than the fastest human alive but it still can't fold a towel.

Join me for a fast, myth-busting tour of AI humanoid robotics, world models, and the real gap between human and machine. Expect live thought experiments, hard numbers, and a look at what's coming next from emerging fields like World Models — the technology Yann LeCun just left Meta and raised $1.03 billion to build.

**[Callum Pender](https://www.linkedin.com/in/callumpender/), Machine Learning Engineer @ Faculty**

**Identifying terrorist propaganda**

The Faculty team partnered with the Home Office to build a model that helps smaller online content platforms automatically detect Daesh propaganda videos. The talk will explore the use case, the solution and the impact of this tool and how Faculty approaches technical deployment in highly secure domains.

Callum Pender is a Machine Learning Engineer at Faculty with a wealth of experience working within the National security domain, he has successfully designed, built and deployed AI solutions to solve real world problems across a large number of different clients.

**[Nicholas Greenwood](https://www.linkedin.com/in/n--/),** **Head of Engineering @ Aries Global**

**Agents don't care about your codebase**

You can write kiss, dry and yagni all over the place. You can add a skill for every task. You can have great documentation. You can ask another agent to review the first agent. They still don't care. they just do whatever is simpler.

You need many more static code checks than you think. A pr that doesn't comply with your standards must not pass ci.

I'll cover the repo changes that worked for me: Harsh linters and formatters, static code checks that enforce architecture, clear error messages.

We hope to see you there for another great AI filled evening!

**Sponsors...**

The AI Exchange is brought to you in partnership with[ LinuxRecruit](https://www.linuxrecruit.co.uk/) and[ AWS.](https://aws.amazon.com/)

LinuxRecruit is building the next wave of Tech Unicorns in Europe.

Whether you're hiring or exploring your next opportunity, [get in touch](https://www.linuxrecruit.co.uk/jobs).

Description supplied by the event organiser.