So What’s Going to Occur to Product Administration Anyway?

Expensive subscribers,

I’ve been pondering loads about the way forward for product administration.

Once I discuss to mid-career PMs, I hear the identical frustrations again and again:

1. “I’m on the lookout for a director position however all I see are IC jobs.”

2. “I wish to work for an AI firm however there are only a few PM roles.”

3. “I’m too drained to construct at night time after a day of again to again conferences.”

In the meantime, AI-native corporations like Anthropic and Cursor are hiring fewer PMs or delaying their first PM rent for so long as doable.

So what precisely is going on, and what are you able to do about it? Listed below are 8 scorching takes on the place I believe the PM position is headed subsequent and why I’m optimistic.

This publish is delivered to you by…HubSpot

HubSpot has put collectively a complete information to supercharge your productiveness at work. This free useful resource consists of:

• Immediate database: 100+ ready-to-use prompts for speedy implementation

• Trade-specific use circumstances: 15+ real-world purposes throughout varied sectors

• Productiveness finest practices: 21 finest practices to 10x your effectivity with AI

+ In-depth sections on emails, content material creation, buyer help, and information evaluation.

Discover Your Information Right here

I’ve interviewed 30+ founders and leaders at AI-native corporations like OpenAI, Anthropic, Cursor, and Replit. All of them consider the identical factor:

Listed below are some direct quotes from leaders:

1. “You desire a comparatively small set of individuals the place nearly everybody you rent is actually, actually good.” — Dario (Anthropic’s CEO)

2. “We now have <30 PMs as a result of we wish to be the mannequin of what it seems prefer to construct an organization on high of AI.” — Nate (OpenAI’s Head of Enterprise Merchandise)

Sadly, I believe which means there can be fewer PM roles at AI-native corporations. It additionally signifies that your FAANG PM job might be not as safe as you suppose it’s. That’s as a result of…

I labored at Meta through the ZIRP years the place a typical product staff had engineers supported by a PM, designer, information scientist, PMM, content material strategist, and extra. However the irony is that:

AI-native corporations don’t need PMs who persist with their lane and suppose:

1. “I’ll let the PMMs market my product.”

2. “I’ll let the researchers discuss to my clients.”

3. “I’m too busy with strategery™️ to dogfood.”

As an alternative, these corporations need T-shaped builders — folks with deep experience in a single space and sufficient breadth to contribute to many others. Ryo (Head of Design at Cursor) put it finest:

My buddy Ravi made this nice level:

Because of AI (simply strive writing a spec with Cursor), designers and engineers now have extra bandwidth to cowl PM duties like summarizing buyer insights and drafting product technique. That’s why corporations are delaying hiring PMs.

Within the AI period, a PM who can’t construct and an engineer who can’t outline necessities will each battle. In the event you’re nonetheless working at an organization the place you’re advised to “persist with your lane,” then it’s most likely time for a change.

The extra skilled I get as a PM, the much less persistence I’ve for PM theater like:

1. Spending extra time documenting what you’ll ship than really delivery.

2. Obsessing over inside docs as an alternative of dogfooding the tip person product.

3. Making choices by committee and compromising the product alongside the way in which.

Firms and groups that may construct within the morning and get person suggestions by lunch will win over these nonetheless doing waterfall growth.

I don’t care in the event you’re a PM director with 10+ years of expertise. You must nonetheless be taught to spin up a prototype shortly, present it to actual clients, and iterate.

This isn’t simply my opinion. “Quick beats proper” (Claire Vo) and “execution beats technique” (Ami Vora) are direct quotes from product leaders whom I respect. Group-led growth and constructing speedy suggestions loops are what matter.

Hopefully, I’ve satisfied you now that constructing is vital. Now your first intuition could be to spend $1,000+ on an AI course to “be taught to construct.” Simply maintain this in thoughts:

I cringe once I see folks publish: “I simply bought my AI PM certificates from XYZ course!” That is the equal of placing “AI PM” in your LinkedIn:

No one cares that you simply bought an AI course certificates. What hiring managers wish to see is proof of labor:

AI corporations are express about this. Right here’s a quote from Anthropic’s careers web page:

The one credential that issues to AI-native groups is what you’ve really constructed and what tangible concepts you must enhance their product.

And the one method to be taught these items is to get your fingers soiled, make errors, and begin delivery. Watch my AI tutorials in the event you want an additional push.

At this level, chances are you’ll be pondering: “Look Peter, I’m simply too drained after back-to-back conferences and placing my youngsters to mattress to construct.”

I hear you, I’m in the identical boat. Once I was younger and naive, I believed “I actually have to change into a PM so I can lastly construct good merchandise!” However right here’s the irony:

As an alternative of constructing, you spend extra time in evaluations, stakeholder conferences, and efficiency calibrations. In the event you’re not cautious, you possibly can simply lose contact with the small print and the craft of creating the product.

So what’s the answer? For me, it’s being very intentional about defending my time by:

1. Auditing recurring conferences. I ask: “Does this assembly nonetheless have to exist?”

2. Batching conferences. I attempt to cluster conferences so I’ve mornings for deep work.

3. Defaulting to async. I’ve change into superb at async Slack threads.

This doesn’t all the time work and will backfire. However it’s a trade-off that I’ve consciously made in order that I’ve extra time to suppose, dogfood my product, and iterate with AI.

The humorous factor is that I’ve had a number of CPOs e-mail me after leaving their jobs asking the best way to vibe code. I additionally know a number of PM pals who’ve proactively moved from management roles at non-AI corporations to IC roles at AI-native corporations in order that they will construct once more.

That’s as a result of for the primary time ever, ICs who use AI instruments can have extra leverage than folks managers. Nikhyl (former Meta VP) has a terrific quote about this:

Don’t let your single-minded pursuit of climbing the ladder take you away from crafting the product if that’s what you really take pleasure in doing.

Right here’s one other uncomfortable fact:

Nikhyl has a terrific framework for fascinated about completely different work intensities:

1. 996 (9 am – 9 pm, 6 days per week): AI startups the place you’re married to work. There aren’t any boundaries—work is precedence zero, every part else comes after. These startups are incompatible with parenting younger youngsters, caregiving, or well being points.

2. On-line after hours: The norm for many PM jobs at FAANG and development corporations. You’re on Slack at 9 pm however can set up some boundaries. Depth is available in waves (e.g., launch durations, incidents).

3. No nights and weekends: Co-workers log off at 6 pm and there aren’t any weekend expectations. The trade-off is often slower development and fewer cutting-edge work, however you’ll have extra time to construct by yourself.

It’s a must to be trustworthy about your work and life constraints. I’ll by no means work at a 996 firm as a result of I wish to watch my youngsters develop up. I do work nights and weekends generally as a result of I care about delivery good merchandise.

However these days are the exception, not the norm.

I simply wrote 6 factors about how vital it’s to be taught to construct, however I additionally suppose that:

Aligning folks is tough and stays beneficial even on a small staff.

AI can draft your launch plan, however it might’t stroll into a gathering and resolve a impasse between sturdy personalities. It might probably summarize person suggestions, however it might’t learn the room when a beta flops and you must determine whether or not to delay.

The PMs who thrive will mix builder velocity with affect — they will each present the long run with a working app, after which carry the staff alongside for the trip.

Consider constructing as a method that will help you align with different folks. A prototype offers folks a a lot better sense of your product than any slide deck.

The PMs who thrive on this new world can have what I name “determine it out” power — the willingness to put on a number of hats and clear up issues as an alternative of ready for somebody to show you.

This shift is tough for large tech PMs as a result of most of us have been educated to work inside clearly outlined roles and processes. We’ve discovered to not “step on toes” in order that stakeholders can provide us good suggestions throughout evaluations. We accumulate FAANG credentials like infinity stones (I’m responsible of this). We’ve discovered to test packing containers and stack credentials our whole lives.

It’s time to get up.

In an AI-native staff, hesitation is extra harmful than stepping on toes, iteration velocity issues greater than excellent planning, and proof of labor issues than credentials.

With the ability to simply determine it out is what issues.

The unhealthy information: There’ll probably be fewer nice PM jobs.

The excellent news (and why I’m optimistic): The PM jobs that stay can be extra enjoyable — faster-moving, larger affect, and nearer to crafting the product.

As PMs, it’s as much as all of us to evolve our position towards what we really love doing. I’ll depart you with three tangible subsequent steps:

1. Study to construct one thing your self. Begin by watching my newbie tutorials to construct a Chrome extension in Cursor, lovely web sites utilizing Replit, and AI headshot and film discovery apps with Claude Code. However don’t simply watch the tutorials — construct them your self!

2. Tighten your suggestions loops. Inform your staff that you simply’re now not doing waterfall growth. Shorten the time from thought → shipped change → real-world studying to someday cycles. Reject doing greater than 3 rounds of inside iteration with out speaking to a single buyer. Observe community-led product growth.

3. Be trustworthy about your constraints. Working 996 for an AI native firm comes with actual sacrifices. Search for corporations the place you are able to do good work and dash for brief durations however nonetheless have time to be with your loved ones.

The PM career is at an inflection level. The talents that bought many people right here gained’t take us the place we have to go.

However in the event you construct the proper abilities and produce that “simply determine it out” power, you may quickly be doing essentially the most fulfilling and impactful product work of your profession.