ep. 72. The Tale of the Bluenose: A Holiday Reflection on Automation
4 min read
As a Canadian living in the United States, this week always feels like a pause between holidays - Canada Day on July 1, and Independence Day on July 4. So today’s edition is a short one, written from the East Coast of Canada.
Yesterday, I visited the Fisheries Museum of the Atlantic in Lunenburg, Nova Scotia. A docent told the story of the Bluenose, the legendary Canadian schooner that now graces our dime.
It starts in 1920, with the America’s Cup, which was abruptly postponed due to high winds. Spectators retreated below deck. Conditions were declared unsafe. East Coast fishermen scoffed.
These, they said, were “pink tea affairs.” If a bit of wind was enough to send elite sailors indoors, they argued, maybe it was time real seafarers stepped in. And so, from that moment of derision, the idea for the International Fishermen’s Cup was born: a test not of racing yachts, but of working fishing schooners.
In December 1920, construction began on the Bluenose - a fast, elegant, and fully operational fishing schooner. She was launched just a few months later, on March 26, 1921, in Lunenburg.
That October, Bluenose entered the International Fishermen’s Cup off the coast of Halifax. The race carried a strong sense of national pride, drawing large crowds and making front-page news in both Canada and the US.
It was a best-of-three series, and Bluenose won the first two races with ease. Bluenose then went on to defeat the American challenger Elsie, winning the International Fishermen's Trophy for Nova Scotia.
The schooner went on to win the cup in every subsequent competition, capturing front-page headlines across Canada and consistently delivering impressive fishing hauls between races. Its captain, Angus Walters, was invited to meet King George V of England during the 1935 Silver Jubilee - a testament to the vessel’s growing renown.
However, despite her legendary status, tides were shifting in the fishing industry.
Fresh fish had overtaken salt cod as the dominant industry in the North Atlantic, driving demand for faster vessels. Motorized schooners and trawlers were starting to take the place of traditional sailing vessels. They didn’t rely on the wind - or the skill of seasoned sailors. Captain Walters recognized the shift and even attempted to retrofit the Bluenose with an engine. But the era of sailing schooners was ending.
During World War II, the Bluenose was docked in Lunenburg and later sold in 1942 to the West Indies Trading Company. Converted into a coastal freighter, it was wrecked off Haiti in 1946 and abandoned. In 1963, a replica, Bluenose II, was built in Lunenburg using the original plans. It now serves as a tourism ambassador.
The story of the Bluenose left me thinking about our current state of automation, largely driven by advances in AI.
What is our Bluenose? As trawlers and motorized vessels replaced sailboats, an entire skillset and ethos were devalued. What ‘undefeated’ practice is now on the cusp of being obsolete?
How do we navigate the in-between? Captain Walters tried to adapt - he retrofitted the Bluenose with an engine, but it wasn’t enough to compete with changing markets. How do we sequence our automation strategy?
Under what conditions does automation succeed? The vessels that would replace the schooners delivered more value at scale. Automation often succeeds when value and technical feasibility align, enabling scale. Yet, as we’ll see in the last article in the section below, that alignment is too often missing in today’s AI deployments.
Human-Computer Interaction News
AI and R&D Productivity: This McKinsey article covered AI’s potential to accelerate R&D productivity, which has been declining over time. These include generating more (and more creative) design candidates, evaluating those designs faster via AI surrogate models, and streamlining research operations, including automating documentation tasks.
Exploring the Risks of AI in Mental Health Care: A new study investigated the use of LLMs to replace mental health providers. Compared to trained human therapists, LLMs (e.g., GPT-4o) provided problematic guidance, including expressing stigma toward those with mental health conditions, and responding inappropriately, such as encouraging clients' delusional thinking. How might trained therapists work with LLMs, augmenting rather than replacing their expertise?
What Workers Want from AI vs. What Startups Are Building: This Stanford study investigates whether AI startups are automating the right tasks by surveying nearly 2,000 workers across over 7,000 tasks, evaluating both worker desire and AI capability for automation. It finds a major mismatch: most tasks workers want automated are not currently being targeted by AI tools. 41% of Y-Combinator startups are targeting low-desire, low automation tasks. This may help explain why only 25% of CEOs feel AI have delivered expected ROI. AI systems need to deliver on both the desirability and capability axes to be successful.
Building an emerging product in XR, AI, or Robotics? Sendfull helps you get to market faster - and with greater confidence - using neuroscience-backed evaluation frameworks. Reach out at hello@sendfull.com
That’s a wrap 🌯 . More human-computer interaction news from Sendfull in two weeks!