Apple Mail and Notes: Which AI features actually ship, and what they cost
Brandi Reed · Sep 14, 2026Most of Apple's AI in Mail and Notes is already on your device. The catch is hardware, not a subscription.
Tech News
AI vendor contract red flags: what to audit before signing
Brandi Reed · Sep 13, 2026The demo went well. The contract is where the deal …
Workflow
Germany's military is testing 30 AI tools. The real story is the procurement, not the algorithm
Marta Kowalczyk · Sep 11, 2026A 2027 prototype target and a public snub of Palantir's …
Wellbeing
Three documented ways to keep AI out of your evenings
Hiroshi Nakamura · Aug 28, 2026Named studies back a short list of boundary practices — …
Tools
Is ChatGPT's free plan enough for your actual workday?
Marta Kowalczyk · Aug 26, 2026Per OpenAI's own pricing page, the free tier covers casual …
What a context window is, and why long AI chats start losing the thread
Marta Kowalczyk · Aug 17, 2026Vendor documentation says the window holds your prompt, the model's …
Running local models with Ollama: what your laptop can honestly handle
Marta Kowalczyk · Jul 22, 2026Ollama made local AI a one-line install — the honest …
Workflow
What ChatGPT, Claude, and Gemini actually remember about you, per their own documentation
Rekha Patel · Aug 21, 2026Each vendor now ships a memory feature that carries context between chats. Here's what each one stores, what it doesn't, and how to turn it off, according to OpenAI, Anthropic, and Google's help pages as of August 2026.
A verification workflow for AI outputs: tier the claims, then check the load-bearing ones
Rekha Patel · Jul 26, 2026"Trust but verify" is a slogan; a workflow is tiers — which claims need checking, against what source, at which moment — and the design spends your skepticism where mistakes actually cost.
Apps
What Granola's meeting notes actually cost, and where your audio goes
Brandi Reed · Aug 12, 2026The AI notetaker just hit a $1.5 billion valuation. Here's what the free, Business, and Enterprise tiers include, and what its privacy policy says happens to your recordings.
AI baked into browsers: Edge Copilot, Chrome's assistants, and what to enable
Brandi Reed · Jul 30, 2026The browser became an AI surface — summaries, page Q&A, tab help, agentic browsing — and the settings that matter concern what the assistant reads and when it acts.
What one interruption really costs, now that your AI tools ping you too
The published field figure is 25 minutes and 26 seconds, not the 23 minutes the internet repeats — and scheduled AI notifications are a new source of the same problem.
Audit what you offload to AI: the quarterly hour that keeps you sharp
Delegation is the point of these tools — but unexamined offloading quietly atrophies the skills you'll need when the tool is wrong, absent, or inappropriate, and a simple quarterly audit keeps the trade deliberate.
Screen eye strain: what clinical sources actually recommend
Digital eye strain is real, common, and temporary — and the evidence-backed fixes are boring: the 20-20-20 rule, blink-awareness, proper distance and lighting, and an up-to-date prescription.
GPT-5 shipped — what actually changes at your desk
OpenAI's August 2025 launch replaced model-picking with automatic routing and rationed its deepest reasoning; here's the overlooked part.
What the EU AI Act's code of practice actually requires from AI model makers
A voluntary rulebook now sits behind the EU AI Act's toughest deadlines for OpenAI, Google, Anthropic and the rest. Here's what it actually asks of them, and who's still holding out.
Why small AI models are suddenly everywhere
The frontier gets the headlines while the small models quietly take the work — economics, latency, and on-device privacy explain the shift, and vendors' own routing announcements document it.
Is AI actually taking entry-level jobs? Here's what the data says
A new Stanford payroll-data study, NPR's reporting on it, and Gallup's engagement numbers each answer a different piece of the question — and none of them agree on the whole picture.
Why updates matter: the patching habit CISA keeps asking for
Most compromises run through known, already-patched vulnerabilities — the documented pattern behind security agencies' endless patching advice — and the habit that closes it is small, automatic, and boring.
Laptop security on the road: the airport-and-hotel threat model
Travel device security is about physical proximity, unfamiliar networks, and border crossings — three threats home hygiene doesn't face, with documented countermeasures for each.