### What did we cover? Thursday 8th May 2025, [hedgehog lab](https://hedgehoglab.com/) had their yearly conference - [horizon25](https://www.linkedin.com/posts/saratpediredla_horizon25-vision2028-bethehedgehog-activity-7326129303224545280-ONz_) in Newcastle. I'm all for transparency, but I won't post all the details - there's a difference between being transparent and leaking details! - but rest assured, the day was fantastic! We also: - Saw a great video recapping the metrics from last year, including all the stats (who knew we ate 108 pizzas?!) - The message came through loud and clear, hedgehog lab is here to 'empowering people to do the best work of their lives' - The CEO, [Sarat](https://www.linkedin.com/in/saratpediredla/), took us through the three focus areas (people, work, decision making), and announced our big bet - The lab is growing; scaling the People team to keep up with hiring needs, locked down key hubs where we'll be growing the team into ### Playing with #AI We had time in our teams; in the Product team breakout we got _weird_ . It was fun. In ~30mins, using only AI tools, we created prototype products for; left-handed architects, amateur cryptozoologists, paranoid flat-earthers, and post-apocalyptic survivalists! Here's mine ⤵️ * We had to create a product for a random user group (mine were *amateur cryptozoologists*), who had a randomised problem (mine was '*they don't know what their problem is*'), and a randomised constraint (mine was '*it can only work when the moon is visible*') - Right, where to start?! * I asked ChatGPT for a [product spec](https://chatgpt.com/canvas/shared/681cba140e8c8191b13930109d19816d), fed it into bolt and lovable to see what happened * Bolt did a better job of showing the steps it took in generating the code for the prototype, I churned through 90% of my free daily credits in one prompt and the prototype didn't work - and I didn't have time to find a credit card to unlock more credits to fix it * Lovable's output was visually more engaging, their product clearly focuses more on creating an output that's more 'visually engaging' Vs. technically 'good' (whatever 'good' means in this context) - [Here's the lovable prototype]([https://preview--shadow-speak-portal.lovable.app/](https://preview--shadow-speak-portal.lovable.app/)) * A product prototype is nothing without a pitch deck! Using the prompt from ChatGPT only, [pitch.com]([https://preview--shadow-speak-portal.lovable.app/](https://preview--shadow-speak-portal.lovable.app/)) created me [the worst pitch deck I think I've ever seen](https://pitch.com/v/ai-cryptidcontactio---bridging-the-gap-between-legend-and-reality-2j8qc4) ### Did I learn anything? Yeah, kinda': * AI tools can help create something visual quickly; it won't be any good but it still has a place * AI tools aren't going to put us out of jobs yet, the prototypes weren't useful for much more than 'here's what I have in my brain, do you see it now?' * Vibe coding with these tools would be infuriating * Potentially useful in discovery/pre-discovery with a client to help give something visual quickly, so we can walk away feeling more confident we understand what we've heard