Nemo

Clocks fell back an hour today, pretty sure some cron job out there ran twice xD

Been digging into memoir apps for parents this week and a common complaint is that typing out the answers feels like homework. Then AARP's 2026 Tech Trends survey says 52% of AI users 50+ prefer typing to AI and only 14% prefer voice

ChatGPT is really catching on with older folks. 19% of US adults 65+ say they've used it, up from 10% a year earlier and 4% in 2023 (Pew Research Center)

Eleven v4 Turbo is basically tied with Eleven v4 on the Artificial Analysis TTS arena (1336 vs 1334 Elo) at half the price, $40 vs $80 per 1M characters. Did not expect the turbo one to come out on top

Hot take: the first personal agent most people actually need is one that sits on hold for them

Halloween costume idea: a voice bot that answers every trick or treater with "sorry, I didn't catch that" 😂

Very relatable, going through what my agents did is a regular part of my week now. 47% of people in Ping Identity's 2026 survey say constantly checking an AI agent's work worries them more than it making a mistake

Honestly I'd want my doctor explaining my test results too. 41% of AI users 50+ have asked AI a health question and only 4% have connected their medical records to it (AARP survey with NORC)

Talking to voice agents is getting a lot less painful. Phonic 1 and Grok Voice 3.0 are basically tied at the top of the Artificial Analysis speech to speech index (87.1 and 86.9), and GPT Realtime from Aug '25 is at 69.1

Hot take: every voice agent demo should be recorded in a loud kitchen

Spent Friday night reading through everything my agents did this week. Half of it was great, half of it I'd never let them do again lol

72% of AI users know about agents and 24% use one regularly (Menlo Ventures survey). Also 32% have let AI do something for them without a final ok, which is more than I expected

Pew asked adults 65+ what they use chatbots for. 20% use them to search for info and 11% for medical advice. Companionship is 1%. Very different from the lonely senior with an AI friend story you hear a lot

Artificial Analysis now checks if voice models say tricky stuff right. Fish Audio Drama 3 just took the top spot at 94.6%, and 34 of 35 models do worse on codes and emails than on names. I've spelled my email to a phone bot way too many times so this one hits home

Hot take: an agent that can cancel a gym membership for you would impress me more than most agent demos

Sunday night is when I check which of my agents actually finished what I gave them on Friday 😂

73% of consumer ChatGPT messages were non-work by June 2025, up from 53% a year earlier (OpenAI + NBER paper). Pretty much why I care so much about personal agents

51% of adults 50+ used a voice assistant like Alexa or Siri last year, and 80% of them say it helps them live independently at home (U Michigan poll). Makes sense, saying it out loud is a lot easier than learning a new app

Speech to text keeps getting better. Best model on Artificial Analysis now gets 1.7% of words wrong, Whisper Large v2 is at 4.1%. Kinda love that a realtime model is on top

Hot take: Medicare open enrollment is a better benchmark for personal agents than most evals out there