Dear forecaster

People come to askaforecaster.com, and magic ball tells us all. Because they are willing to pay for an answer, questions tend to be interesting, and I’ve found it very meaningful to answer them. The below has been lightly edited for anonymity and published with the clients' permission.

For $150: Will AI eat my product design job?

Dear forecaster,

Current AI tools are augmenting work for engineers at many technology companies. The same thing is happening for non-engineering product-related roles as well – for instance, Product Designers are having AI included in design tools to create interface mock-ups faster; Product Managers are using LLMs to write drafts of briefs and memos; etc. Presumably, as these tools improve employee efficiency, companies will need to hire fewer and fewer employees to get the same amount of output.

Specifically for these non-engineering product-related roles (which, you may have guessed, I fall into), I wanted a quick forecasting take on how employer demand for these roles may fall in the next 5 years. Is it more likely to expect a modest 10% fewer positions? Is a 30-50% reduction possible? An industry-shattering 80-90%? And what factors might make a position more or less resilient?

Best,

Denis McDesigner

Dear Denis,

We can think in general terms about what happens when you automate a fraction of a job away, and then go through the case of product designers specifically

What happens as you automate larger fractions of a job away

• Initially as you automate some fraction of a job away, labor is still necessary for the rest, and so automation is “labor augmenting”. This is probably the case for 1-20% of a job.
• As you automate more and more of a job, possibly you might get a Jevons paradox, where increased efficiency and decreasing price leads to much more demand, because you unlock new usecases. This was the case with coal during the industrial revolution, where cheaper coal led to much more demand for coal for new industries.
• Alternatively, you do reach market saturation, jobs are automated away, and so the remaining workers compete for lower spots. For instance, automatic checkouts didn’t lead to more checkout lines, it led to fewer checkout lines and automatic checkout lines being overseen by fewer workers.
• As you automate the last 1%, the remaining labor goes from being still fairly valuable because it unlocks the automated ediffice, to being worthless because it’s worse than fully automating something away. The CEO of a company with 1000 automated employees might be really valuable, because it unlocks the value of those employees… until the CEO position also becomes automated.

Some cases that may give you something to think about:

• Horses. Initially the steam engine/trains automated long distance travel, but increased the number of last-mile trips needed, so the population of horses continued to grow through the 1800s.
• Chess. There was a brief period where humans+chess engines were better than chess engines alone, but it passed pretty fast.
• Coal. Making it cheaper created much more demand for it.
• Doctors. By gating medicine prescription behind a visit to the doctor rather than a talk with chatgpt, they retain a role. Joining a cartel or a cartel-like organization might be a good option?

What might happen in the case of product designers/nontechnical roles in general?

To the above story we can add:

• Friction. Society is organized one way, and organizations may be slow to change, to understand how to replace employees, etc. Yes slower companies may get outcompeted by leaner ones eventually, but the change might take time.
• Taste paradox. AIs can produce designs for your boss, but will your boss be able to judge whether they are good or not? Presumably this is what Apple had when Steve Jobs was around but now has lost. If taste can be operationalized it can be automated away, but this may be hard because striving too hard is also tacky, and because trends are also cyclical. Your mileage may vary.

With this in mind, over the next five years, for product designers in particular:

• A slight increase seems possible
• A modest 10% fewer positions, or a 30-50% reduction both seem in the cards.
• An industry-shattering 80%-90% seems very unlikely. Translators have only gone down a modest 9% since 2019 (2024, 2019), and their wages have gone up slightly (though not adjusted for inflation).
• If there is some reduction, it might manifest through not hiring junior designers, rather than firing senior product designers.

I have less to say about product managers, partly because I understand what they actually do less. Some thoughts:

• They seem more disintermediatable in principle
• They seem like a combination of two jobs:
  • making things happen, oversight, coordination; reducing friction
  • a political position that’s not quite useful but also hard to dislodge.
• With more time, maybe worth keeping an eye on BLS statistics (https://data.bls.gov/oes/#/industry/000000) and see how different types of managers (e.g., “Computer and Information Systems Managers”) change with time.

What should you do to position yourself in this situation

• Put yourself in a position where automating your job away has a large upside for you, e.g., in a startup that does this where you get some equity.
• Emphasize the parts of your job that are not automatable.
• Collaborate with startups to offer the fraction of your jobs that they need, e.g., a product design review rather than a full design which they can vibecode internally.
• A constraint that is salient to me these days is distribution. Maybe post more on social media?

Best,
Nuño

For $150: When will Canva IPO? (2025-05-22)

Dear forecaster,

I’d like an estimate of when Canva will IPO. Ideally your best point estimate, and a distribution over dates. If it isn’t possible to do this justice within the $150 price point, let me know.

Best,

Angel Investorovich

Dear Angel,

Here are our probabilities:

• 2025: 7%
• 2026: 25%
• 2027: 19%
• 2028: 15%
• 2029: 8%
• Not by 2030: 26%

You can see our much longer writeup here. In short, we look at a few baserates, particularly dropouts from the population of unicorns, and time to IPO after hiring a related Chief Financial Officer, and we update on recent news that they are intending to do so. Overall forecasters found the question engaging and fun to think about :)

Best,
Nuño & the Samotsvety team

For $20: What will $NVDA price be at EOY 2025

Dear forecaster

What will $NVDA price be at EOY 2025? The question is self-explanatory. Feel free to give a probability weight to each price target or answer however you see fit. Thanks :)

Best,

Daisy McTrader

Dear Daisy,

The $20 answer to your question looks as follows:

• <100: 29.43%
• 100-160: 18.83%
• 160-200: 10.04%
• 200-280: 14.54%
• 280-340: 6.92%
• 340-400: 5.29%
• >400: 14.93%

I got this by running this utility from my friend Jonathan Mann that predicts a stock price by taking a monte carlo simulation of its past stock price changes. That is, it draws future price changes from the distribution of past price changes.

Things I like about the above forecast:

• Good baseline
• Conceptually neat
• Captures that NVIDA is a volatile stock
• Subjectively captures that it still has room to go up

Things I don’t like about the above forecast:

• Does NVIDIA really have a 3x upside still? Much harder now than a few years ago now that multiple competitors trying to eat its lunch
• You’re not going to beat efficient markets with this approach
• Anchors too much on “NVIDIA is good at going up" 
• Does not look at fundamentals

Best,
Nuño