GenAI Transforms Labor. Google’s new quantum chip. Europe's AI factories.
GenAI Agentic Systems change how we assess market niches by transforming labor. Google's quantum chip is promising yet distant from practical utility. EU's AI factory gamble to stimulate innovation.
Labor Transformation Due To The Rise Of GenAI Agentic Systems👷♂️
In earlier editions, we explored how GenAI agentic systems challenge traditional SaaS incumbents like Salesforce and whether they can succeed. Today, let's consider how they can influence the labor market in the future.
📊 Expanding Addressable Market Size
With the past generations of software, the market size was estimated based on the revenue potential of analog or digital predecessors. These estimates guided venture investments, assuming multiple-X returns on such investments.
GenAI agentic systems are breaking this pattern. For the first time in digital transformation history, these tools are ready to revamp legacy software or marginally improve the automation of part of a business process, automate a complex process, and replace the labor in one go. This dramatically expands the addressable market size, including:
Software costs being replaced.
Labor costs being reduced
👉 Economic reality is more complex and nuanced: While individual compensation packages may inflate, the total cost of labor will be shrinking in areas of impact, freeing up significant budgets. Additionally, professional service costs (e.g., legal or accounting) most likely will experience deflationary impacts.
🏗️ Unlocking Budgets Through Labor Market Gaps
Another opportunity lies in addressing unrealized funding caused by labor shortages. Many industries, including the public sector, face high demand for specialized labor but lack sufficient supply due to:
Low level of compensation
Extremely specific candidate requirements
Governments, especially under the new U.S. administration, may reallocate funding toward solving these gaps. Elon Musk predicts up to $2 trillion in government spending— maybe a fraction will be invested in niche agentic systems.
Many of these roles fall under what David Graeber describes as a “bullshit job”— a job where even the person doing it secretly believes the job shouldn't exist. The genuine hope could be that GenAI agentic systems could lift these morally exhausting routine tasks from people. Of course, if they continue to exist after upcoming waves of deregulation.
💡 Pricing for Outcomes, Not Seats
As noted in a previous newsletter, GenAI services tend to be priced based on customer outcomes, avoiding the traditional cost-per-seat model. This pricing structure aligns more with the incentive to replace inefficient and unproductive jobs.
🌍 Enabling Social and Economic Shifts
In parallel reality (read: Japan), there is also an experiment to reduce the number of working days, giving people more time in hopes of fixing the population decline. By increasing overall productivity, agentic systems could make this shift either more feasible, urgent, or even mandatory soon.
👉 Imagine a world where work gets done more efficiently, and people have more time for themselves (or children). Unfortunately, at this moment, we only see mainly drops in job openings, and it may or may not be connected to the rise of AI.
Google’s new quantum chip wears the willow
Last week, Google announced Willow, its new, “state-of-art”, 105-qubit superconducting chip with which it’s demonstrated an error-corrected surface code qubit as well as a new, bigger quantum supremacy experiment based on a Random Circuit Sampling chip. It produced a round of applause from the media and even a pat on the back from Elon Musk. Wow!
Quantum computing experts, physics, and enthusiasts began to chew on the press release.
In his blog post, Scott Aaronson shares his perspective on the announcement and makes some interesting remarks about Willow’s computation results validation.
“Google’s quantum computation would take ~10^25 years for a classical computer to simulate; it would also take ~10^25 years for a classical computer to directly verify the quantum computer’s results. For this reason, all validation of Google’s new supremacy experiment is indirect, based on extrapolations from smaller circuits”
Sabine Hossenfelder brings us all back to earth with highlights on the real-world applications and classic computers always catching up:
It's exactly the same calculation that they did in 2019 on a ca 50 qubit chip. In case you didn't follow that, Google's 2019 quantum supremacy claim was questioned by IBM pretty much as soon as the claim was made and a few years later a group said they did it on a conventional computer in a similar time.
( . . . ) the consequences for everyday life are zero. Estimates say that we will need about 1 million qubits for practically useful applications and we're still about 1 million qubits away from that.
Also, it's been a recurring story ( . . . ) because some other group finds a clever way to do it on a conventional computer after all.
And, of course, the grand quantum computer skeptic still didn’t find enough time to analyze the latest claims but still agitates not to trust Google on any such due to the flaws in the past ones. We are waiting, Gil!
We did not study yet these particular claims by Google Quantum AI, but my general conclusion apply to them “Google Quantum AI’s claims (including published ones) should be approached with caution, particularly those of an extraordinary nature. These claims may stem from significant methodological errors and, as such, may reflect the researchers’ expectations more than objective scientific reality.” (Our specific contention points are relevant to Google’s newer supremacy experiments but not directly to the quantum error-correction experiment.)
Also, it seems like the quantum threat gradually becomes a bit less of a problem. This year, the milestone of 13% of global TLS 1.3 traffic already using post-quantum encryption was reached, with 62.1% of all websites adopting the latest protocol version.
But who cares if the market is happy?
Europe tries to figure out AI factories
Apologies, we could not resist covering this bit of news in a bit more detail. AI infrastructure, data and people are the areas a 1.6B EUR government investment that will land into 15 different locations across Europe. Although the proposal authors are still trying to figure out what an AI factory actually is (and if there's any chance of actually providing market-competitive services to local startups),
This picture looks quite grim and looks like just a half-hearted attempt and hasty effort to do something in response to the recent report by Mario Draghi.
If you’d prefer a summary of the situation around innovation in the EU, check out the post by Andrew McAfee.
“The US has a large and variegated population of valuable young from-scratch companies. The EU simply doesn’t. The American population of arrivistes worth at least $10B is collectively worth almost $30 trillion dollars — almost 70 times as much as its EU equivalent.”
We will be keeping our eye on how the story about AI factories evolves and will be back to it in one of our following editions.
🔑 Things you probably already read last week
Facebook’s Red Book: A great example of corporate culture carved in stone and printed as a beautiful book
Cloudflare’s 2024 year in review: internet traffic grew 17.2%, overall 41.3% mobile, Starlink traffic 3.3x, 6.5% of traffic considered to be malicious.
Stablecoins, Decentralization, AI assistants and AI-enabled local business as key 2025 investment topics from YC and a16z