The ELECTE blog
Back to the front page All articles
Newsletter
The Strawberry Problem
"How many 'r's' in strawberry?" - GPT-4o answers "two," a six-year-old knows it is three. The problem is tokenization: the model sees [str][aw][berry], not letters. OpenAI didn't solve it with o1-it got around it by teaching the model to "think before you speak." Result: 83% vs. 13% in Math Olympiad, but 30 seconds instead of 3 and triple the cost. Language models are extraordinary probabilistic tools-but you still need a human to count.AI StrategyThe Future of Construction and Real Estate: A Lesson from the Construction Health Sector
Why do generic AI solutions fail in construction? AI that does not distinguish "load-bearing walls" from "partitions" is dangerous. The results of specialized AI: -68% design errors, -31% inventory, -28% construction duration. Royal London Asset Management: 708% ROI, -59% energy consumption. Construction is second to last in digitization among industries-ideal ground for demonstrating the value of vertical versus generic AI.AI StrategyThe ROI of AI implementation in 2025: comprehensive guide with real case studies
$3.70 return for every dollar invested in AI-the top performers come in at $10.30. But 42% of companies have abandoned most projects by 2025, citing unclear costs and uncertain value. Novo Nordisk: 12 weeks to 10 minutes for clinical reports. PayPal: -11% fraud losses. 74% achieve positive ROI within first year, but only 6% become "AI high performers." The question is not "can we afford AI?"-it's "can we afford to delay?"AI StrategyThe AI Productivity Paradox: thinking before acting
"We see AI everywhere except in productivity statistics"-Solow's paradox repeats itself 40 years later. McKinsey 2025: 92% of companies will increase AI investments, but only 1% have a "mature" implementation. 67% report that at least one initiative has reduced overall productivity. The solution is no longer technology, but understanding the organizational context: capability mapping, flow redesign, adaptation metrics. The right question is not "how much have we automated?" but "how effectively?"AI StrategyThe Efficiency Paradox: Does AI Make Us More Stupid?
Socrates feared that writing would destroy memory. He was right-the storytellers who recited the Iliad have disappeared. But we have gained the ability to preserve ideas on a global scale. Every technology "de-trains" something and enhances something else. AI is no different, but it is faster: 5 years instead of 20-30. The question is not "does AI make us stupid?" (no), but "are we aware enough to choose what to delegate and what to keep trained?"AI StrategyThe Model Context Protocol (MCP): a new "USB-C" for AI that transforms business workflows
"USB-C for AI integrations"-that's what they call the Model Context Protocol, and OpenAI, Google, Microsoft and Amazon are all adopting it. Over 1,000 MCP servers created by the community in just a few months. But serious vulnerabilities have emerged in 2025: prompt injection, silent "rug pull" of definitions, credential exposure. Gartner warns that authentication is still limited. The promise is huge: a universal language to connect AI to any system. The advice: non-critical pilot projects, cautious curiosity instead of hasty adoption.NewsletterThe Great Deception: Why AI Understands Emotions Better Than It Admits
82% AI vs. 56% human accuracy in emotional intelligence tests-the Geneva and Bern study demolished our last reassuring myth. ChatGPT-4 not only outperforms humans in existing tests-it creates new ones indistinguishable from those of professional psychologists. Microexpressions, speech analysis, contextual understanding-AI reads emotions we ourselves do not recognize. The question is no longer "can it understand emotions?" but "how do we harness this superior understanding while keeping human values at the core?"AI StrategyThe Business of the Good Old Days: nostalgia as a competitive advantage
While OpenAI and Anthropic still seek sustainable business models, MyHeritage and FaceApp print money by improving photos from the 1990s. The inconvenient truth: Consumers pay more to improve the past than to imagine the future. It's the "20-Year Nostalgia Cycle" monetized by AI at the perfect time-degraded digital archives + technology to restore them + generation with buying power. $17B→$50B market by 2030. But if we optimize only to look backward, who will invent the future?Page 36 of 39