Products often degrade when revenue pressures replace user focus, a process dubbed 'enshittification'. The article breaks down how this creep shows up, in LinkedIn’s paid reach, Duolingo’s gamified streaks, and more, then offers concrete tactics for product teams to push back, from championing user‑centric metrics to building internal firewalls against short‑term profit grabs.
As AI narrows the gap between concept and code, companies are pulling creative technologists out of marketing labs into the heart of product development. The article traces the role’s evolution, from ad agency oddity to a hybrid of design, engineering, and storytelling, showing how today’s product teams rely on them to prototype, iterate, and launch at speed.
Airline mobile apps are shedding static home screens for context‑aware flows that surface gate info, baggage drop, and real‑time alerts exactly when travelers need them. The shift, highlighted by recent UX research, promises smoother airport journeys and could boost loyalty by cutting frustration at the most critical moments.
Perplexity’s latest UI lets you prompt the chatbot for hex‑code color suggestions, turning its text‑first engine into a quick diverging palette generator for data visualizations. By blending two sequential schemes around a neutral midpoint, designers can produce balanced, accessibility‑ready colour ramps without manual trial‑and‑error, and the workflow even syncs with Adobe Color for seamless export.
Matt Turnbull told recently laid‑off Xbox employees to use AI chatbots like ChatGPT for resume drafting and emotional coping, saying the tools can reduce mental strain. Creatives slammed the advice, fearing AI threatens their jobs and seeing it as tone‑deaf amid massive Microsoft layoffs.
Andrew Chen repurposes the micromort concept to invent tongue‑in‑cheek business gauges, like Lies per Second and Meetings per Decision Ratio. By turning lies and meetings into quantifiable rates, leaders can pinpoint how much misinformation and meeting overload is draining productivity, and set concrete targets to cut waste.
CostPerPrompt aggregates live pricing for over 230 LLM models and turns token rates into monthly cost estimates. Its calculators factor in prompt caching and batch discounts, so product teams can accurately forecast API spend before committing. The site updates automatically, cutting guesswork from budgeting AI features.
The Wage Against the Machine dashboard compares the hourly cost of human labor across countries with the price of completing the same tasks using leading LLMs. It pinpoints knowledge‑work tasks that are now cheaper to automate, hinting at imminent shifts in hiring and pricing strategies.
Pronto Stream offers a real‑time data wire that connects AI agents to over 70,000 live feeds, from sensors to social streams. Agents can subscribe instantly, turning raw streams into actionable decisions without building custom pipelines. The service promises lower latency and scalable observability for any AI‑driven product.
Symbio runs locally on Apple Silicon and closes the feedback loop by generating outputs, critiquing them, and fine‑tuning a LoRA adapter on‑the‑fly, turning each correction into a dedicated skill. This eliminates cloud dependence and human‑in‑the‑loop overhead, letting teams build adaptable, private AI assistants.
Evidence-to-Skill inserts a quarantine layer between external data and AI agent skills, automatically validating and attributing inputs to block prompt‑injection and data‑poisoning attacks. The GitHub project ships a tiny, dependency‑free auditor that records evidence sources and audit reports, but it only flags known patterns, human review remains essential.
Crew is a local‑first IDE that lets you and any number of LLM agents edit code side‑by‑side, share a folder, and run up to ten concurrent agent threads without any cloud service. By keeping everything on your machine it preserves privacy, avoids API costs, and lets teams experiment with AI‑driven workflows instantly.
You’re caught up.
That was all 12 stories in the 2 August issue.