NYTimes: “The outlook for software developers is more likely evolution than extinction, according to experienced software engineers, industry analysts and academics. For decades, better tools have automated some coding tasks, but the demand for software and the people who make it has only increased. A.I., they say, will accelerate that trend and level up the art and craft of software design. “The skills software developers need will change significantly, but A.I. will not eliminate the need for them,” said Arnal Dayaratna, an analyst at IDC, a technology research firm. “Not anytime soon anyway.””
WSJ: ““I believe that travel is a sport . . . and it takes training and the right equipment.” Thus saith the Points Guy. As Warren Buffett is to value investors so Brian Kelly is to those who treasure airline, hotel and credit-card points. Fans flock to his website for tips and deals, and now they have his book, “How to Win at Travel,” a primer in the calculus of rewards-points accumulation. “Points are real, nonliquid assets,” writes Mr. Kelly. “While the IRS doesn’t generally consider points as income and won’t tax you on them (in most cases), points are an extremely worthwhile currency.” And more than that, “credit cards are a part of being American”—the average U.S. consumer, we are told, has 3.84 credit cards.”
Hollis Robbins: “The AGI systems launching now can reason, learn, and solve problems across all domains, at or above human level…in the AGI era, the only defensible reason for universities to remain in operation is to offer students an opportunity to learn from faculty whose expertise surpasses current AI. Nothing else makes sense.” [via Arnold Kling]
Dan Shipper: “Language models transform text in the following ways: Compression: They compress a big prompt into a short response, Expansion: They expand a short prompt into a long response, and Translation: They convert a prompt in one form into a response in another form. These are manifestations of their outward behavior. From there, we can infer a property of their psychology—the underlying thinking process that creates their behavior: Remixing: They mix two or more texts (or learned representations of texts) together and interpolate between them.”
Tanay Jaipuria and Charles Rubenfeld: “Achieving GPT‑4–level performance now costs nearly 1,000 times less than it did just 18 months ago. Similarly, Deepseek’s R‑1 model has slashed the cost of delivering entry‑level intelligence by almost 27 times in just three to four months…This significant drop in costs will enable companies to integrate AI into the free tiers of their products, potentially unlocking AI for over a billion new users and sparking innovative applications in everyday software. Looking ahead, two key factors will sustain this trend: the development of increasingly smaller models and the shift toward on‑device inference, driven by advances in chips and infrastructure that allow models to run efficiently on computers, smartphones, and other devices.”
