The AI Playground 2024: 10 ways to get hands-on experience

🧠 Knowledge Series #54: Build dashboards, create songs, conduct deep research with Claude, Perplexity, Suno and more.

🔒The Knowledge Series is available for paid subscribers. Get full ongoing access to 50+ explainers and tutorials to grow your technical knowledge at work. New guides added every month.

---

Hi everyone 👋,

As 2024 draws to a close and some of you look forward to having a well deserved break, when the excitement of eating, drinking and celebrations inevitably dies down a little you may find yourself with some free time and perhaps even feeling a little bored at times (maybe that’s just me 😬).

And what better way to cure that boredom than by playing with AI tools to help you to get some hands-on experience before the inevitable return to office in January?

In this Knowledge Series, I’m going to share some practical (and hopefully fun!) ways you can get some hands-on experience with some of the world’s leading AI tools from 2024. So that by the time you’re back in the office, you’re fully up to speed and equipped with some new practical AI skills.

We’ll explore some of the top AI products from the year including some personal favorites such as Perplexity, Claude, Suno, Gemini, Runway, Copilot and others.

Coming up:

1. Write your own song to celebrate your team’s achievements at work this year with Suno

2. Build a data dashboard in Claude

3. Build a market research assistant with Google’s new Deep Research product

4. Create your own pomodoro timer in Claude Artifacts

5. Build a research collection in Perplexity

6. Create a branding video using generative AI and Runway

7. Build a basic web app in Replit

8. Design a slide deck using Gemini or Copilot

9. Troubleshoot a product you’re using with Perplexity

10. Teach yourself how to read API documentation using Perplexity and NotebookLM

---

The Knowledge Series

1. Write your own song to celebrate your team’s achievements this year with Suno

OK, this one is a little left field and probably isn’t something you’d use regularly use at work. But it’s a helpful exercise in understanding just how far generative AI has come in such a short space of time.

Using a tool called Suno, you can now generate entire songs based on a single prompt. I did this the other day to wish my sister a happy birthday and she was delighted but also slightly terrified by it.

Suno is free to use to create a few songs so you'll be able to get yourself some hands-on experience on their free tier. In our example, we’re going to imagine that we want to send our colleagues or team mates an upbeat song to celebrate all of the things we’ve achieved together this year.

Here’s the prompt we can give Suno to do this:

an upbeat, 90s inspired song which celebrates my team's achievements this year. including:

1. releasing a new mobile app
2. hitting Q4 sales targets
3. upgrading our content management system

In response from Suno, we get two alternative versions of the 90s inspired pop track which it calls Victory Lap, along with lyrics and a shareable video.

Here’s the shareable audio lyric video it produces for us:

You can use whatever musical style you want in the prompt and when paired with some visuals, this kind of thing could be a great way to celebrate the end of the year. Including people’s names is a fun way to make it more personal, too.

2. Build a data dashboard with Claude Artifacts

One of the biggest developments in AI this year was the release of new capabilities which allow non-engineers to build simple apps. GitHub released Spark, Poe released Previews and OpenAI gave us Canvas. But Claude Artifacts seems to have emerged as an early leader in this space of building light apps that can be used day to day.

Dashboards are part of our everyday lives and work and while it used to take a data analyst or engineer a few days to gather requirements and build the dashboards you need, now you can do this in a few minutes with Claude Artifacts.

Using the following prompt along with a CSV of the data you want to visualise, you can build your own data dashboard in seconds (in my case I’ve uploaded a spreadsheet with some Google Analytics data):

create a data visualization dashboard artifact using the spreadsheet attached. I'd like it to visualize all key metrics.

After Claude gives you its first attempt you can then converse with it to move data points to where you need them and re-jig any styles. For example:

put the total sales, total visitors and average bounce rate metrics on the same row

Within just a few minutes we now have a dashboard which is available on the web to share with colleagues:

If you wanted to iterate on this, Claude lets you go further by adding interactive elements like date pickers or other functional buttons, too.

3. Build a smart market research assistant with Gemini Deep Research

It’s been a mixed year for Google but during the last quarter they’ve managed to redeem themselves slightly with the release of products like NotebookLM which finally led to a breakthrough into the mainstream and by simplifying their AI offerings all under the Gemini brand.

This week, as we mentioned in the briefing, the company revealed its latest new major product release: Gemini 2.0. The next iteration comes with upgrades across various parts of the Gemini AI stack but one product in particular that stood out as being particularly helpful for product teams is its new research tool, Deep Research.

Deep Research allows users to ask Gemini to perform research on a user's behalf and but crucially, it uses what it learns throughout the research process to continuously evolve and refine its results. Once finished, the results can be exported into a Google Doc.

Let’s imagine we’re planning a new mobile app launch targeting gen z users and we want to understand the payment landscape and what payment methods might be suitable for this group of users as part of our product strategy.

Here’s a video example which shows me using Deep Research, including the prompt, research outline and Google Docs / Sheets export that it creates for us: