Last week of February, eh! That was fast… This week we start with Alice Hamilton which fought against industrial lead poisoning long before the world was ready to listen, proving that progress often means battling resistance until hindsight catches up. Meanwhile, the Gorton typeface, a silent workhorse of urban design, shows how even the most overlooked details shape our daily lives. On the darker side of technology, the loneliness epidemic is fuelling GenAI-powered romance scams, turning emotional vulnerability into a billion-dollar fraud industry – surprise, surprise! And speaking of illusions, chatbots are now the digital age psychics, convincing users of intelligence where there is only statistical prediction. But not all these new ventures are cautionary tales! For example, Jeff, the Tetris-playing AI, is a proof to ingenuity, even if it can still be bested by something as simple as sparkly visual effects – really!
In this issue
5 piecesAlice Hamilton, a pioneering physician and toxicologist , spearheaded early 20th-century efforts to expose the dangers of industrial lead poisoning
. Investigating hazardous workplace conditions, she confronted corporate resistance and government inertia, yet her findings led to significant public health reforms. As Harvard’s first female faculty member, she challenged gender barriers while advocating for safer labour conditions. Her warnings about leaded gasoline and industrial toxins, initially dismissed, were later vindicated, reinforcing her status as a foundational figure in occupational health and safety.
Alice was the kind of person history often overlooks until decades later, when everyone realises she was right all along. Imagine spending your life uncovering scientific truths about toxic workplaces, only to be dismissed by powerful industries and ignored by policymakers - sounds familiar? So, whether it’s climate change, environmental toxins, or workplace safety, the pattern remains: an underdog fights for reform, industries resist, and eventually, after irreversible damage, society admits, “Oops, we should’ve listened.” Hamilton’s work is a reminder that progress isn’t just about discovery; it’s about persistence, public pressure, and - unfortunately - waiting for the world to catch up, and these next 4 years are going to be incredible...
Marcin Wichary's article, " The Hardest Working Font in Manhattan ," goes into the history and ubiquity of the Gorton typeface
- a 150-year-old font often overlooked despite its widespread presence. Originating in the late 19th century at the British company Taylor, Taylor & Hobson, Gorton was designed for engraving machinery and has since become a staple in industrial settings, adorning everything from control panels to public signage. Its distinctive, utilitarian design has made it a silent workhorse in urban environments, particularly in cities like New York.
Isn't it fascinating how some of the most pervasive elements in our daily lives go unnoticed? Gorton is a prime example; a typeface that's silently guided us through subway stations, elevator panels, and industrial machinery, yet remains largely anonymous. Its unassuming, mechanical design speaks to an era where functionality trumped aesthetics, embodying the grit and resilience of urban life. For me, discovering Gorton's story feels like unearthing a hidden layer of the city's identity, reminding us that even the most mundane details have rich histories waiting to be explored.
The rise of loneliness is not just a social issue - it’s a security crisis. This article explores how increasing isolation makes people more vulnerable to romance scams ,
which have already cost victims billions of dollars. As scammers adopt generative AI to automate and scale their fraudulent activities, the threat is growing. These scams manipulate victims emotionally, often mimicking tactics used by domestic abusers, making them especially difficult to combat. Experts warn that AI-driven scams will become more sophisticated, making it even harder for victims to recognise deception.
What a wild irony; technology connects us more than ever, yet loneliness is at an all-time high. Is like the movie "Her".... But the fact that scammers are capitalising on this human vulnerability isn’t surprising - at least for me -, but the AI-powered scale of it is unsettling. What’s even scarier is that the same AI we use to craft dating profiles or flirt in DMs is being used to deceive and exploit. The line between romance and fraud is getting blurrier. Maybe the bigger issue isn’t just AI - it’s that as a society, we’ve created a world where loneliness is so rampant that millions are willing to send money to a stranger just to feel connected. That’s the real crisis.
In " The LLMentalist Effect ," Baldur Bjarnason argues that chat-based Large Language Models (LLMs) create an illusion of intelligence similar to techniques used by psychics.
He suggests that users, especially those enthusiastic about AI, may experience subjective validation, interpreting generic responses as personally meaningful. This effect is amplified by the hype surrounding AI, leading users to perceive LLMs as more intelligent than they are. Bjarnason cautions against overestimating LLMs' capabilities, comparing reliance on them to consulting a psychic for advice.
It's fascinating and a little unsettling, how easily our minds assign 'intelligence' to a machine that’s just predicting the next word in a sequence. This comparison to psychic cons is quite sharp, reminding us that the real trick isn’t in the LLM itself, but in our willingness to find meaning where there may be none. After all it's just math. But whilst these models have their uses, matching them with genuine understanding is a mistake. The GenAI industry thrives on hype, but critical evaluation is crucial, without it, we risk slipping into a world where statistical patterns are mistaken for true insight. Maybe it’s time we balance our fascination with a healthy dose of scepticism, ensuring these tools serve us without pulling the wool over our eyes.
In " First Place in Tetris 99 Using Computer Vision, Classical AI, and a Whole Lot of Free Time ," the author details the development of "Jeff," an AI designed to autonomously play Tetris 99 on the Nintendo Switch.
Jeff utilizes computer vision to interpret the game state, a depth-first search algorithm with a custom utility function to determine optimal block placements, and a microcontroller to input moves via USB. This setup enabled Jeff to consistently rank in the top 15 players and occasionally secure first place.
Oh boy! This is just na interesting journey, isn't it? This AI Tetris 99 is both a technical marvel and a clear reminder of AI’s limitations. The fusion of computer vision, algorithmic strategy, and hardware tinkering showcases the sheer dedication of enthusiasts who push the boundaries of what machines can do. Yet, even a highly optimised AI can be tripped up by something as simple as a sparkly visual effect; proof that unpredictability isn’t just a human trait. As AI advances, solving these issues will be crucial for autonomous systems navigating fast-paced, visually complex applications—whether in gaming or beyond. Out of example; Microsoft is investing a lot into this area recently - gaming, I mean.
Last Week in AI
- Hello, DeepSeek Open Infra! | github.com
- All of Humane's AI pins will stop working in 10 days | engadget.com
- Microsoft’s Xbox AI era starts with a model that can generate gameplay: from The Verge | theverge.com
- Perplexity claims to have purged Chinese censorship and propaganda from its new DeepSeek clone | sherwood.news
- Nvidia helps launch AI platform for teaching American Sign Language | venturebeat.com
- Alibaba shares jump 14% after pledge to invest ‘aggressively’ in AI; from the FT | ft.com