I predicted AI (kind of)

I predicted AI (kind of)

Every day I feel like I’m coming across a new AI tool that completely boggles my mind. Today’s discovery was Rewind.ai - it’s insane 🤯 I’ve recently been absolutely in love with Raycast for the past several months, but I’ll definitely be digging into Rewind this weekend.

Now that AI models have been released into the wild, people are building insane products so quickly, that even a year or two ago, many of us could never have imagined any of this would be possible in just a couple years.

Back in 2021, I took a Product Management Certification course with Brainstation, and the final assignment was to present really any product idea to the rest of the group. The idea I did was a really dumbed down version of my original idea\, only because it wouldn’t have been super fleshed out. At the time, if I were to have said: “Using AI ✨… “ I would have probably pretty quickly lost my audience’s attention.

So, what was my presentation? Here’s the recorded slide deck 👇🏽

Pretty meh, right?

I was honestly more proud of the slides and the janky logo I made than the actual product 😅

My actual idea was way cooler. But before I get into it, if you didn’t watch the video, just skip to the 3rd slide and read my problem statement. Basically, I just think that it’s weird that tech companies don’t have a “Product Dictionary” of sorts that lets anyone, at any time within the company know how any single functionality works on the surface level, and if they want, the actual code. The complexity of the product changes this for sure, but it was just weird to me that I’ve had to ask several people questions that should realistically be accessible by searching somewhere. Oftentimes, there’s questions that I’ve asked that no one is actually 100% confident in, so we end up having to spend the time to actually go in the product and test it.

The current solution to this is either finding the right employee who is a genius and/or has tribal knowledge, or more commonly done, sifting through a knowledge base or community that’s never truly up to date because it requires way too many resources to maintain. Usually, you won’t even find all the edge cases or granular functionality questions that you were looking for.

So, my solution?

AI ✨

The farthest extension of my idea was basically: LLM’s that are trained on your codebase as well as all your documentation, both internal and external, providing you the ability to ask questions and receive answers that are literally based on the code itself.

The documentation piece is already do-able and I’ve seen companies implementing this by having models read through all their written resources. Workbounce.com is one example. Next, is the code part. I’m sure there are teams actively building this, but I could imagine the difficulty and potentially the lack of ability of current models to sift through complex codebases with a lot of different references across many different files.

“The current easiest way to get the answers you need is to send a message to a colleague or a whole channel and pass off the cognitive burden to them.” - Adam Smith, Workbounce co-founder

Instead of asking one of your colleagues: “Hey Jared, do you know if when users click the ‘pay now’ button, does it send an API call to Stripe immediately? Or do we store some information on our servers beforehand?”, you could ask AI and it would tell you exactly what it does, with the associated code, and maybe even the UI of before and after the pay now button. This is the future of product development and all connected teams. Even customer facing roles.

I was just looking at one of my old blog posts and realized that I predicted this:

intelligent AI powered searching softwares that use integrations to search all your various platforms that keep your company’s information. I think that over the next several years, as available APIs continue to expand, and internal platform-based integrations extend, AI powered search tools will be comonplace in the workplace. - What we can learn from product teams

Excited for my future relationship with Mr/Mrs GPT. 🤖

Btw, if you are interested in taking the PM course at Brainstation, here are the questions you should ask yourself before taking any online course:

I think that the value that online courses provide are based on a these key variables:
  1. Quality of educators & access to them
  2. 👉🏽 Who’s teaching you? Do you just sit and listen to them or is it interactive?

  3. The content
  4. 👉🏽 Is the content you’re being given unique?

  5. The status associated to certification company
  6. 👉🏽 Will people recognize the name of the certifying organization?

  7. The other students
  8. 👉🏽 Do you work with the other students at all? Will you build your network?

My answers:

  1. High quality & pretty interactive.
  2. No - it’s moreso the perspective of the instructors that was really valuable.
  3. Yea, I’d say so. There’s only a handful of companies that have awareness in the PM world, and they change every few years it seems too.
  4. Yes! I’m connected with a handful of them and have loved keeping in touch with some of them.

Thank you for making it this far! Please give me feedback. Writing is hard.

Contact Me