Product management in an AI world
Learn how to find and hire rainbow unicorn pegasuses in one easy step!

Product management in an AI world

And lo, the utility of LLMs!

Read the article on Substack .

If your Large Language Model of choice cannot accomplish a task you set for it, wait a couple of months. A new, bigger, more capable model will arrive and make child’s play of your ask. I don’t know if it’s reasonable to say AI is progressing faster than expected. I suppose that depends on who you ask.

Benchmarks published since OpenAI’s release of GPT-4o show marked improvement compared to a year ago. Notably, all foundational models show similar increases, and closed-source models outperform open-weight models because of, aherm, money. Lots and lots of money. OpenAI has raised $11.3B to Mistral’s $113M, a curiously symmetrical 100x fundraising lead.?

That amount of money doesn’t, however, buy OpenAI a model that is 100x better than those produced by their less well-funded competitors.?

Yes, LLMs are still lazy, petulant teenagers and struggle with basic tasks. Their promise doesn’t yet match the hype. But they are also expert tutors, software developers, and capable interns.

AI in product management

AI is changing product management in two important ways.

  1. Use LLMs to build domain expertise quickly.
  2. Any product manager can now build software prototypes and designs.

Building domain expertise

In the past, product leaders had to choose between domain expertise and product management expertise when building their product organizations. Those days are gone.

Imagine you are writing the job description for a product manager. You need a seasoned veteran with 7-10 years of hands-on product management experience, a track record of bringing products from concept to retirement, and a background in public key infrastructure and cryptography. You send the job description to your recruiter, who responds, “Rainbow unicorn pegasus.”

While you appreciate the pithy sarcasm, you feel compelled to impress upon her the need to find a matching candidate as quickly as possible. All your hopes, dreams, and plans depend upon this person.

After several weeks of searching, you have a stack of resumes in two piles: experienced product managers and people who know PKI and cryptography. There is exactly zero overlap between the two groups. The recruiter ended your latest Slack exchange with a simple “I told you so,” followed by a mic drop emoji.

Instead of relaxing with your partner on a Friday evening, you puzzle over whether teaching a new hire how to do product management or become proficient in PKI and cryptography will be easier.?

Fast forward to today.

In about thirty minutes, I built a GPT to tutor my new product manager to an advanced proficiency level with PKI and cryptography...

Check out the prompts I used to build the GPT and Read the rest of this article on Substack .

要查看或添加评论,请登录

社区洞察

其他会员也浏览了