Subscribe to Making Sense

No spam, no jargon. Unsubscribe anytime.
Makes Sense. You're subscribed to the AI signal for Canadian business leaders. Every two weeks. No hype, no jargon.
All issues Issue 12 · September 17 2026

Making Sense.

AI, translated for Canadian business leaders.

Editor's Note

We're in an interesting moment for AI, where organizations and governments are racing to spark adoption, while members of the public and AI companies themselves echo concerns with the acceleration of AI. The reality is that there is already an enormous amount of latent capability with the technology released today. No matter how much AI model development slows, a long runway of opportunities awaits most companies to explore.

Our advice for now is to keep focused on adapting your business to operate within the new AI paradigms (in any case, they're here to stay), while digging into the new technological possibilities that are on the table. The only constant is going to be change, and trying to wait for absolute clarity is likely to just leave you further behind.

In this issue
  1. 01
    Three stories from the past two weeks.
    Canada's National AI Literacy Initiative launches next week, leaders inside the AI industry call for slower development, and a Québec firm aims to take on the American cloud giants.
  2. 02
    Release Radar.
    OpenAI ships its most capable model yet, Anthropic cuts the price of agentic work, Thomson Reuters builds its own model on decades of legal and tax content, and Google makes video analysis cheaper.
  3. 03
    Four numbers for your next meeting.
    The growth capital gap, the size of Canada's data centre market, how fast OpenAI's ad business grew, and what Codex did for 1Password's engineers.
01 — In The News

Three stories from the past two weeks.

Canada's National AI Literacy Initiative launches next week, leaders inside the AI industry call for slower development, and a Québec firm aims to take on the American cloud giants.

01

National AI Literacy Initiative starting September 21.

Students working on laptops at desks in a bright library classroom

The federal government launches its National AI Literacy Initiative next week, a $13-million program delivering free AI training in three streams: one for post-secondary students, one for educators, and one for the general public. The student course is the first to be released on September 21, 2026, with courses for the general public releasing later in the year.

The program is being delivered by the Alberta Machine Intelligence Institute (Amii), one of Canada's three national AI research institutes, and is modelled on literacy courses Amii has run for the past decade. The curriculum is about assessing AI-generated information and spotting misinformation, bias, and privacy problems. AI Minister Evan Solomon said at the launch that "the best way to ensure success at a time of change and innovation is through education."

Read more →
02

Anthropic, OpenAI, and Musk all called for a slower AI race.

A city traffic light showing amber against an office building

Anthropic CEO Dario Amodei published an essay this month arguing that model capability is outrunning researchers' ability to understand and control it. He proposed three fixes: independent evaluators with employee-like access inside the leading labs, shared safety standards across democratic countries, and eventual international limits on the most dangerous capabilities. OpenAI CEO Sam Altman endorsed Amodei's essay and committed OpenAI to the independent evaluator piece. Elon Musk, founder of AI venture xAI, also agreed. Prior to Amodei's essay, more than 100 companies including OpenAI, Anthropic, and Microsoft signed an open letter warning that AI-enabled cyberattacks will become "far more widespread and sophisticated" as models improve.

Read more →
03

Three US firms hold 85% of Canada's cloud. A Canadian company wants to change that.

An aerial view of a Canadian downtown skyline lit up at night

Three American firms, Amazon, Microsoft, and Google, currently control 85% of Canadian cloud services, according to the Global Media and Internet Concentration Project at Carleton University. Micrologic, a Québec provider of cybersecurity, IT, and data and AI hosting, pitches itself as the alternative. CEO Stéphane Garneau estimates that companies choosing Micrologic over US providers could keep $2 billion in Québec over five years. This month it raised $45 million from two Québec backers, the labour-sponsored Fonds de Solidarité FTQ and a fund managed by the province's investment arm.

Canadian telecoms Bell and Telus have both launched sovereign cloud and AI data centre builds, including Bell's $1.7-billion site near Regina and Telus's Vancouver and Kamloops cluster. But the big three's 85% leaves only 15% for Bell, Telus, Micrologic, and every other provider combined, and the newest Bell and Telus builds are not yet online.

Read more →
02 — Release Radar

Recent launches.

  • • OpenAI ships GPT-6 Astra, its most capable model to date, with state-of-the-art results on computer use, coding, cybersecurity, and science.
  • • Anthropic's Claude Fable 5.1 costs about 25% less than Fable 5 on typical workloads and up to 45% less on agentic tasks.
  • • Thomson Reuters launched Thomson, its own model family trained on decades of proprietary legal, tax, and regulatory content. On the company's own testing, it narrowly beat GPT-5.4 and Claude Sonnet 5 on completeness and factuality in its domains.
  • • Google brought agentic video understanding to its Flash models, improving accuracy on video analysis while cutting token usage and cost substantially.
03 — Quick Facts

Four numbers for your next meeting.

  • • Over the past decade, 33% of Canadian growth rounds were led by Canadian investors, against 74% of US growth rounds led by US investors. PitchBook, via RBC
  • • Canada's data centre market is worth about $14 billion USD and is projected to more than double by 2031. Mordor Intelligence, via BetaKit
  • • OpenAI's advertising business reached $1 billion in annualized revenue run rate about 200 days after launch, with ads now live in more than 40 countries. CNBC
  • • Toronto's 1Password increased engineering productivity by 21% using OpenAI's Codex, building new features and internal tools to production readiness without loosening its security policies. OpenAI