Subscribe to Antonym newsletter at antonym.substack.com and BN Edition at bnedition.substack.com for more.
++ Update after Jan 2024 Agencyhackers The Robots Are Coming++
Slides from Agencyhackers presentation.
Antony Mayfield - Agencyhackers - robots are coming .pdf
Bonus content
Briefing on how models like ChatGPT Premium (GPT 4) are outperforming specialist databases when used with sophisticated prompting techniques. Also the briefing paper we prepared for our clients about how GPT4 (ChatGPT Premium) can beat specialist AIs. It also has a link to the original “Medprompt” paper from Microsoft.
Superprompts briefing paper - Brilliant Noise 12.23.pdf
<aside>
💡 Thank you for attending the Brilliant Noise webinar on how to help your team explore the possibilities of generative AI.
To find out how we can help your organisation contact us at [email protected].
</aside>
Video
https://vimeo.com/845953360
Tools
The AI tools we mentioned using were:
- ChatGPT (of course)
- Poe. Use different AI models and create simple bots.
- Notion. Note taking and project management app with AI built in.
- Lex. A free (for now) online word processor with powerful AI features.
- MidJourney. The best AI image generator right now.
- Consensus. Ask questions and an AI will draw on academic papers for answers.
- Summarize.tech. Put in the URL of a long YouTube video and it will summarise it in detail.
- Goblin.Tools. Designed for neurodivergent people to help organise tasks and other things – but incredibly useful for anyone when they aren’t feeling focused.
Team day agenda
Find out more
To get help running team days, experimentation and capability programmes about generative artificial intelligence complete the short questionnaire below:
https://brilliantnoise.typeform.com/to/Fv3SCRlc
Articles
This is how we structured a team session to help us understand and explore generative AI.
1. Before
A week before the event we asked everyone to start recording any tasks or projects that either they thought AI might be used to improve or just tasks that were difficult and seemed to take longer than they expected.
2. AI-B-C
Kicking off the session we had a presentation on generative AI basics:
- The history of the field
- How the technology works (at a high level)
- Current debates about the AI
- Tips for using it well (questions, awareness of limitations)
- Security and safety advice
- Strengths and weaknesses of AI
3. Demos
- We demonstrated some tools and how some colleagues had used them in their work.
- Showed interesting and practical examples in the pubic realm.
- Showed techniques like giving ChatGPT a “role” to play before asking questions.
4. Exercise
-
Ideally you want an hour or two for this, but we were pressed for time and had a very productive session in one hour and then had pairs or groups present their experience back to the group.
-
Form a small team.
-
Brainstorm ideas about tasks or projects to look at.
-
Choose one project and write a mini-brief:
- What is the goal of the task?
- What are the steps to complete it?
- What would a good outcome look like?
-
Look for opportunities where AI could be used to:
- Summarise complex data or documents
- Carry out research
- Analyse unstructured data like interview transcripts or videos
- Help generate ideas
- Create a first draft of a document
- Take on the role of an expert or coach to give advice about how to approach a challenge
- Create visual illustrations like mood boards
- Improve the quality of writing or communications
(This isn’t an exhaustive list of things you can do with AI.)
5. Sharing and next steps
In the final session each group gave a brief account of:
- The task or challenge they set
- What they experienced in designing their process
- How they felt at different stages of the process (there were a lot of surprises and insights about how to organise ourselves and what AI tools could help with).