Pseudocode activity
Pseudocode is human-readable, not designed for machines/IDEs to use/convert. Usually a mix of plain English and some very simple coding concepts e.g. conditionals/loops.
Write a simple application in pseudocode for:
- A dice-roller app
- A guess-the-number game
- A number sorting program
- A to-do list manager
- An assignment grader
- A word/letter counter
- Choosing a venue for drinks
- Deciding where to go on holiday
- Finding your location in Melbourne based on visible landmarks
DIY Machine Learning Program
Individually or in groups, complete the following. Document everything — perhaps as a pitch to develop an app? — and upload to your student blogs, or your studio SharePoint folder.
Find a problem:
- What do you want to predict?
- What information will you use?
Collect data:
- List the types of information you’ll gather
- Describe how you’ll get this information
Train the computer/write your program:
- What/how will the program will learn from the data?
- Describe any patters the computer might look for
- Write out your program in simple pseudocode
Make predictions:
- How will the computer use what it learned?
- What will it predict?
Check how well it works:
- How will you know if the predictions are good?
- What would you do if the predictions aren’t accurate enough?
Use your program:
- Describe how people might use this in real life
Grasping the everyday
The ‘everyday’ is a key concept for this studio. It comes from the work of Michel de Certeau. Read actively through the set reading from de Certeau’s book, and complete the following:
- Go back through your notes/highlights
- Draw out any key points – copy to a note if you haven’t already
- What are the systems within which you/we operate? What strategies are deployed by those systems to order behaviour/use/consumption?
- What are some of the ways you ‘resist’ on a day to day basis, i.e. your everyday tactics?
Workflows activity
Map out the following workflows — multiple samples for each.
- Daily workflows, e.g. morning routine, getting ready for bed, making a cup of tea/coffee
- Uni/study workflows, e.g. completing a class group activity, approaching an assignment, keeping track of study/learning/ideas, career development.
- Research workflows, e.g. studying a topic, compiling a quick précis or presentation
Creative workflows
- 500-word narrative (/fragment)
- 2-minute audio piece
- 1-minute video
- Social video (30-60sec)
Each person at the table to choose one of these – ensure you’ve covered all options.
Think about how you would approach – from concept to delivery.
Write out your procedure.
Delegating to AI
- Explore some of the tools listed on aggregator sites like futuretools.io, alternativeto.net or futurepedia.io
- On your workflow/s, where could you delegate tasks to a generative AI tool? Note this somehow – a different colour notation, or a symbol. Be thorough!
- Choose two tasks. Develop a testing scenario to evaluate a tool’s performance at those tasks – what procedure will you deploy? What do you need to run the test? How will you check your results?
- Possible methods that may inform your scenarios/tests are the scientific method and usability testing (UX method).
Testing AI tools with prompts
- Choose three tools from the ‘web-based’ section of the resource list on Canvas
- Develop prompts to complete/respond to the following:
- How might you improve your collaboration skills per the earlier workflow task?
- What are some alternative approaches to begin tackling university assignments?
- How can you improve your research workflow?
- Input your prompts to the three different tools and compare the results (keep prompts and all responses neatly labelled in a word doc)
Prompt tweaking/challenges
- Choose one tool, e.g. ChatGPT. Give it a persona and try one of your prompts again (or start a new chat and give it a different persona – retry your original prompt and note differences)
- Request that the tool present a response in a different form/format.
- Find a prompt tweak/request that the model cannot complete or refuses to do. Why has this occurred?
- Ask a chatbot about other tasks/scenarios where it thinks it could be of use to you. Note these down for further testing/exploration. (You might try one basic prompt for this, and then try again with some extra context about your personality/commitments/preferences, etc)
- Combine all your notes, ideas, thoughts, etc into a single document. Choose a tool that allows you to share documents: share this mega-doc with it, and ask it to generate five possible ideas for this week’s blog post. Note these down for possible development.
Leonardo.Ai Hackathon
- Generate an AI self-portrait. You may only use word prompts: no image guidance.
- Generate a poster for the studio.
- One version can be a background image incorporating elements of the studio to date/what you hope the studio will be.
- One version should attempt to generate text within the image.
- Consider three parts of your creative workflow from earlier. Generate three icons for those tasks or parts of your process.