Site icon Canadian Technology Magazine

Canadian tech and the AGI Shock: Why “AGI Will Freeze the Economy” Matters for Canada

Sofia Alvarez
Canadian tech and the AGI Shock

Canadian tech and the AGI Shock

 

Table of Contents

Outline

Introduction: The AGI thesis and why Canadian tech leaders must care

Matthew Berman’s central thesis is blunt and urgent: when AGI reaches or exceeds human capability in knowledge work, the determinant of economic value will shift from human labour to access to compute. “The only limitation to knowledge work becomes how much compute you have,” Berman explains. That statement cuts straight to the competitive anatomy of the digital economy. For the Canadian tech sector—represented by startups in the GTA, research clusters in Montreal and Edmonton, and the cloud/data-centre investments in Quebec and Alberta—this is not a distant intellectual exercise. It is a design constraint for strategy, workforce development and public policy.

Canadian tech leaders should treat the AGI thesis as both a warning and an opportunity. The warning: unchecked transitions could entrench inequality and lock segments of the population into persistent underemployment. The opportunity: proactive investment in compute, skills and policy could let Canada shape an AI-driven future that benefits a broad cross-section of citizens and firms.

What AGI means: compute, labour displacement and value capture

When the video refers to AGI—artificial general intelligence—it describes systems that match or exceed human performance across a wide range of knowledge tasks: software development, customer service, data analysis, project management, copywriting, legal drafting and more. The crucial economic observation is that once machines can perform knowledge work reliably, the marginal cost of deploying that labour drops to the cost of compute and energy.

Consider the following simplified logic chain:

  1. AGI can replicate knowledge work at or above human levels.
  2. Knowledge work becomes fungible and infinitely replicable so long as compute is available.
  3. Access to compute (servers, GPUs, specialized chips, electricity and data access) becomes the primary scarce resource.
  4. Ownership and control of compute infrastructure translate into disproportionate economic rents.

Translated into market terms, labour’s bargaining power diminishes because human labour can be substituted at scale. Capital—those who own and control compute—accrues the value. The video frames a stark sociopolitical outcome: “Wherever you are socioeconomically at that moment, that’s where you’re going to stay.” This is not purely rhetorical. It’s an economic forecast driven by the changing marginal economics of production in a knowledge-intensive economy.

Early evidence: hiring trends, youth employment and market shifts

There are already empirical signals that labour markets are changing. Notably, a recent Stanford paper titled “Canaries in the Coal Mine” analyzes ADP payroll data through mid-2025 and identifies a roughly 13% relative employment decline among 22–25-year-olds in the most AI-exposed occupations. That is a targeted early-warning sign: entry-level positions—those roles that historically provided on-ramps into careers—are drying up where AI can substitute human effort.

Complementary data from Handshake, a career networking site for students and recent graduates, shows a 15% drop in entry-level postings and a 30% increase in applications per job for the class of 2025. Those two data points together tell a chilling story for new labour market entrants: demand for traditional entry pathways is down while competition for what’s left is intensifying.

Yet the picture is nuanced. Wages have not universally collapsed; older workers in similar roles have seen flat or even growing employment. The immediate effect appears to be about the fraying of the traditional bridge from education to meaningful paid work, rather than a wholesale collapse of all human roles.

What this means for Canadian tech hubs

Canadian tech ecosystems are particularly exposed to these dynamics for several reasons:

For Canadian founders and executives, the takeaway is clear: talent pipelines and compute access are strategic resources that need to be proactively managed.

Compute as the new capital: why data centres and chips matter to Canadian tech

One of the stronger predictions in the thesis is that firms will race to lock down compute capacity. The logic is simple: AGI runs on GPUs and specialized accelerators; the more of that hardware a company controls, the more AGI “workers” it can deploy simultaneously. This has already prompted major global players to invest in massive server farms.

For Canada, compute race dynamics create both leverage and risk. Leverage: Canada can compete to host compute infrastructure by leveraging abundant clean power (Quebec’s hydroelectric resources, British Columbia’s hydro, and growing renewable projects in Alberta), political stability, and data-residency advantages. Risk: if compute concentrates in foreign-owned mega-clouds outside Canadian jurisdiction, domestic firms and public institutions could suffer reduced access or poorer terms.

That is why the compute conversation is not merely technical. It’s a national industrial policy question. Canadian tech stakeholders should evaluate strategies that include public-private partnerships to build local cloud capacity, tax incentives for data-centre construction, and regulatory reforms that ensure the electricity grid can scale with compute demand.

Why cheaper compute might not save workers: Jevons Paradox explained

At first glance, democratization of compute sounds like a silver bullet. If the marginal cost of running AGI falls to the hardware and electricity price, won’t every company—large and small—gain access to cheap AI labour? Not necessarily. Jevons Paradox offers a cautionary counterpoint.

“As some resource becomes more efficient and cheaper to use, overall consumption of that resource tends to increase rather than decrease.”

Applied to compute: as chips and inference become cheaper, new use cases that previously were uneconomic will emerge. Companies will deploy more AI in more parts of their value chain. That expansion can raise aggregate demand for compute, pushing overall spending on compute higher even if the unit price falls.

The implication is that cost declines do not guarantee democratized control. If only a few players have the capital to scale compute at massive levels, they will capture the efficiency and monopolize the economic gains. Cheaper compute increases total usage and therefore increases the scale advantages of incumbents who can absorb variable usage spikes and invest in orchestration, data pipelines, and product integrations.

UBI, UHI and the social safety net debate

Faced with the possibility of structural unemployment or a reduction in labour’s economic value, many technologists and policymakers have turned attention to Universal Basic Income (UBI) or Universal High Income (UHI). Figures such as Sam Altman and Elon Musk have publicly endorsed the idea that a guaranteed cash transfer could stabilize societies during transition.

UBI is simple in concept: regular, unconditional cash payments to citizens to cover basic needs such as food, housing and health care. UHI takes the idea further—payments would be high enough to allow people to choose how to spend their time, supporting creativity, entrepreneurship and non-market activities.

Arguments for UBI/UHI in the AGI context include:

There are also counterarguments familiar in public policy debates: concerns about inflation, political feasibility, moral hazard and the fiscal burden on governments. Yet several policy analysts argue that UBI could incentivize entrepreneurial activity, citing research that shows people with a basic financial floor pursue riskier projects—new businesses and retraining—that have positive social externalities.

How UBI/UHI could play out in Canada

Canada has already experimented with guaranteed income pilots (notably the Ontario pilot cancelled in 2019), and municipal-level experiments continue to attract attention. If AGI-induced transitions accelerate, Canada could become a logical testbed for scaled pilots, leveraging provincial capacity to run cohort-based trials, and federal fiscal levers to fund experiments targeted at tech-displaced workers in cities like Toronto, Vancouver and Montreal.

UBI or UHI does not cancel the need for job creation or upskilling. Rather, it is one component of a larger safety-net and industrial strategy that includes targeted retraining, incentives for demand-side job creation, and measures to democratize compute.

Job augmentation vs automation: where humans remain essential

Not all roles will be eliminated. Influential thinkers at the frontier of AI development suggest that near- to medium-term AGI will often handle the “middle” of tasks while humans continue to handle the ends: prompt, contextualize, verify and deploy outputs.

Balaji Srinivasan (cited in the video) frames this as humans remaining “in the loop”—prompting AI, orchestrating agent swarms, and reviewing output for accuracy, ethics and alignment. This pattern leads to job transformation rather than total automation: the tasks people do will shift, skill sets will adapt, and new roles we cannot yet imagine will emerge. This is consistent with historical technology shocks where occupations evolved instead of disappearing wholesale.

In practical terms, knowledge-worker roles may evolve to emphasize:

The result is a bifurcated landscape. Some jobs—particularly those consisting of repetitive middle-of-the-pipeline tasks—are likely to be automated. Others will change in character and demand higher levels of synthetic, interpretive or creative skill that combine domain knowledge with AI fluency.

Voices from the field: how leaders are interpreting the transition

The video intersperses viewpoints from industry leaders to illustrate the range of interpretations. Canadian tech readers should consider these opinions not as predictions but as strategic signals.

Aaron Levie (CEO, Box)

“The parts of the organization that are using AI will be the ones that get more budget.”

Levie’s point is organizationally pragmatic: AI adoption will concentrate resources on teams that demonstrate increased execution capacity. In other words, companies may redeploy headcount and budget toward AI-enabled efforts because they scale faster and yield measurable efficiencies. For Canadian enterprises, the lesson is to embed AI into core value chains rather than silo it as a pilot program—a shift in resource allocation and corporate governance.

OpenAI leadership perspective

“There is also going to be tremendous job creation, jobs that we can’t even think of today.”

OpenAI’s technologists emphasize that AI will create new categories of work as the technology matures. That aligns with the historical narrative of technological change: new industries and roles emerge even as others decline. For Canadian tech, the challenge is to build ecosystems that can host those emergent industries—places where entrepreneurs can start new companies and scale them efficiently.

Reid Hoffman and demand creation

Reid Hoffman highlights a simple economic mechanism: when production costs fall, total demand often expands into previously impossible categories. Analogous to how cheaper digital tools enabled a flood of new media and content creators, AI will permit entirely new creative and enterprise activities. For Canadian tech, that is an argument to nurture creative AI entrepreneurship in cities across the country.

Balaji Srinivasan on the role of humans

Balaji’s perspective is crucial for workforce planning: many tasks will require human ends-of-chain involvement for the foreseeable future. That means educational curricula and corporate training should emphasize skills that complement AI—strategic thinking, systems design, creative problem solving and human judgement.

What Canadian tech must do: strategy for companies, investors and cities

Given these dynamics, Canadian tech leaders should move from passive concern to strategic action. The following is a pragmatic playbook for stakeholders across the ecosystem.

1. Prioritize compute sovereignty and scale

2. Rebuild entry-level pipelines and apprenticeships

3. Invest in applied AI research and commercialization

4. Reconfigure corporate governance to treat AI as capital

5. Emphasize ethics, verification and quality assurance

Policy prescriptions for Canada: taxation, workforce and social policy

National-level interventions will determine whether Canada becomes a resilient AI economy or a passive host to externally owned compute. The following policy levers should be on the table.

1. Rethink capital taxation and rent capture

If value accrues primarily to owners of compute, fiscal policy must evolve to avoid outsized concentration of rents. Possible approaches include:

2. Scale training with outcome-based financing

Tax credits and subsidies should align with training outcomes. For example, workforce development funds could be contingent on demonstrable job placement in AI-complimentary roles across diverse regions.

3. Pilot universal income schemes

Canada can scale cohort experiments in UBI/UHI targeted at regions disproportionately affected by automation. The federal-provincial structure makes Canada particularly well-suited to run randomized policy pilots and then scale what works.

4. Build a national strategy for compute and data sovereignty

A strategic compute roadmap could identify priority zones for data-centre expansion, grid upgrades and training hubs, making Canada a hospitable jurisdiction for both domestic and foreign investment under defined terms.

Business playbook: how GTA firms and Canadian startups can survive and thrive

Executives and founders should treat AGI as a strategic material risk. A five-point business playbook:

  1. Audit: Map which workflows are AI-exposed and quantify potential value-at-risk and value-at-opportunity.
  2. Invest: Build immediate pilot compute capacity and partnerships with local cloud resellers or co-location facilities.
  3. Upskill: Launch company-wide AI fluency programs with tracked competency milestones.
  4. Reorient: Move budget toward teams that demonstrate AI-enabled productivity gains, but pair them with programs to redeploy affected employees into new roles.
  5. Collaborate: Participate in regional consortia to share training resources, apprenticeships and safety-net design.

For Canadian startups, cheap access to AI tools is a double-edged sword. It lowers barriers to entry but it also intensifies competition. To win, startups must combine AI capability with domain-specific data, customer intimacy and enforcement of regulatory trust—areas where Canadian firms can excel.

Case studies and Canadian examples

While the global players hold significant sway, Canada already demonstrates small-scale wins that can be scaled.

Frequently Asked Questions (FAQ)

Q: What is the core risk AGI poses to the Canadian tech workforce?

A: The primary risk is the shifting source of scarcity from human labour to compute ownership. If AGI enables near-infinite replication of knowledge work, the ability to run AGI at scale—driven by access to GPUs, specialized chips and electricity—becomes the core economic advantage. Without intervention, the outcome could be entrenched socioeconomic immobility for cohorts dependent on entry-level knowledge work.

Q: Will AGI make all Canadian tech jobs disappear?

A: No. AGI is likely to eliminate or transform specific task categories—especially repetitive, middle-market knowledge tasks. However, new jobs will emerge in orchestration, verification, ethics, productization and entirely new industries. The net effect will depend on the pace of automation, policy responses and the ability of Canadian institutions to reskill and create demand in new sectors.

Q: Is universal basic income (UBI) the right answer for Canada?

A: UBI is one tool among many. It can provide a financial floor to stabilize transitions and encourage entrepreneurial risk-taking, but it should be accompanied by training, active labour-market policies, and investments in domestic compute and innovation. Canada is particularly well-placed to pilot and evaluate scaled UBI or UHI models due to its federal-provincial structures.

Q: How should Canadian tech companies prioritize AI investments?

A: Priorities should be strategic: invest in compute capacity where it creates differentiation, integrate AI into core value chains rather than siloed pilots, and pair productivity gains with retraining or redeployment strategies. Treat AI investment like capital infrastructure—plan for lifecycle costs, risks and governance.

Q: Can Canada realistically build sovereign compute?

A: Yes, but it requires coordinated action. Canada has competitive advantages—clean electricity, stable governance and skilled talent—but building sovereign compute at scale needs policy incentives, grid planning and private capital. Public-private models and municipal-level zoning reforms can accelerate buildout in the GTA, Quebec, Alberta and beyond.

Q: What should educational institutions do now?

A: Accelerate AI-fluency programs across disciplines, embed human-in-the-loop training into curricula, expand short-form certifications for mid-career workers, and partner with industry to create paid apprenticeships that expose students to production AI systems. The focus should be on complementary skills: AI orchestration, ethics, domain expertise, and systems thinking.

Q: What immediate actions can municipal leaders take?

A: Municipal leaders can streamline data-centre permitting, create local talent hubs with coworking and training subsidies, and coordinate with provincial utilities to prioritize grid upgrades for strategic compute zones. Cities can also convene consortiums of employers to fund apprenticeships and entry-level placements.

Conclusion: Build compute, protect mobility, design the social contract

Matthew Berman’s central warning—that AGI will freeze economic mobility unless societies act—is neither inevitable nor helplessly deterministic. It is a call to action. Canada’s posture over the next decade will determine whether Canadian tech becomes an inclusive engine of productivity or a class-based producer of concentrated rents.

Three strategic imperatives rise above the noise:

  1. Secure and democratize compute capacity so Canadian firms aren’t disadvantaged by foreign concentration of AI infrastructure.
  2. Invest in human capital that complements AI—training, apprenticeships, and curricula that produce AI-fluent workers.
  3. Design social policies (pilots in UBI or targeted income supports) that stabilize transitions and permit entrepreneurship and retraining at scale.

For leaders in Canadian tech—from Toronto startups to Montreal research labs, from Alberta energy planners to Ottawa policymakers—the task is urgent and actionable. Treat AGI as strategic infrastructure: plan, invest and govern it. If done well, Canadian tech can leverage AGI to create abundance, new industries and a resilient social contract. If ignored, the risk of entrenching a permanent underclass grows real.

Is the Canadian tech sector ready to act? The choices made by business leaders, investors and governments over the next few years will determine whether AGI is an equalizing productivity boon or an accelerant of inequality. Canadian tech has the assets to lead; it needs the strategy, policy coordination and moral clarity to do so.

Engage and respond

How is your organization preparing for an AI-driven future? Are local councils and provincial governments doing enough to secure compute and upskill the workforce? Share strategies, pilot programs or questions for Canadian Technology Magazine’s editorial team and peers in the Canadian tech community.

 

Exit mobile version