Skip to content
Nick Humberstone edited this page Jun 20, 2026 · 4 revisions

SoCraTes UK 2026 Sessions

Recording a session HOWTO

Blog posts

  • add yours here!

Fri 19th June

Harnessing AI by Emily

(Nick)

  • a dog on a lead is a weak harness, it can go anywhere. A carriage and harness on a horse is a good harness, it mostly goes forward maybe veering left or right but always in a scoped direction
  • llm context grows and grows. Md files can help you build harnesses. But if the harness gets too big, even after refining it regularly, you can have subagents that each have their own different harness and then you can avoid amassing too much in a single context. Also break things down into small pieces - it was always good advice, it still is

how to learn new things fast by Nick

(Nick)

  • prompt for the session: news article! Quantum computing has gotten better and means we need to change how we think about some approaches to programming - your boss wants you to look at this - how do you approach it? (I.e. how do you start learning on an unfamiliar topic or unknown domain)
  • things I learnt from people:
  • Anita - a first step of just thinking and writing what questions you have and need to answer to understand it is a good starting point. To make sure you dont go down a rabbit hole, if researching an answer leads to more questions, write these questions down and come back to them later
  • Willem - finding people that know about this area on linkedin and following them could be useful. Researchers are very approachable so if the topic is a niche research thing, ask and they will probably be very happy to share their knowledge. Rss feeds are algorithm free, I should build one for myself!
  • Vlad - researching by using different LLMs gives a different and more rounded set of responses

The way we talk changes the way we work

Led by Ellen, notes here by Nick The name of the session is also the name of a book Ellen had recently started reading. The book talks about how people complaining at a workplace can be reconsidered to show underlying commitments and beliefs people have abiut how they work. And by following this process of analysing the complaint and considering where it comes from we can make progress towards real solutions.

  1. What is your commitment?
  2. How are you responsible for this commitment not being achieved at work?
  3. What are the competing priorities - could include 'i dont want to be seen as the person always challenging things' or 'i want to feel part of the group', not just mechanical calendarisable things
  4. What are the assumptions your making in any of 1-3. (I recognised that an assumption that will be very common for people is 'i am unique in my experience and recognition of this problem'. If something is a problem for you, it is usually a problem for others too)

Capture the flag with Lisi

Using Hackthebox we were accessibly introduced to sec testing and penetration testing and some of the core tools people use to do this. We managed to use sql injection to log in and find a secret on a user's account.

Sat 20th June

Sun 21st June

Previous Editions

Clone this wiki locally