OpenAI Dots: Impressive Debut, Early Backlash

Episode 1 · OpenAI Dots First Impressions: Impressive Capability, Divisive UX (2026-09-30, 22 posts)

OpenAI's new agent product Dots drew hands-on feedback from several high-profile users on day one, with reviews converging on two points: strong AI capability, rough user experience. Matt Shumer, founder of HyperWrite, tested it and said the AI itself is best-in-class with potential to become the world's best agent product, but the UX needs substantial polish before it becomes his daily driver. Early tester Boris MPower was also positive: the specialized dots he tested worked unexpectedly well, and he revealed OpenAI is now expanding access to more users; he called dots the best human-AI collaboration tool he has used to date.

Confirmed

• Matt Shumer (m1/m3, repeated restatement of the same info): top-tier AI quality and big potential, but UX holds it back
• Ethan Mollick (m2): only used it briefly pre-launch, initial impression "quite good," categorizes Dots into the fast-growing Clawlike category (similar to Muse, Grokbot) — assistants built on strong models connected to user data
• Boris MPower (m4/m5): specialized Dots performed "unexpectedly well"; OpenAI is expanding access; the collaboration experience is the best he has used

Why it matters

• Multiple influential users (Shumer, Mollick) gave positive initial takes on day one, and insider feedback corroborates Dots' usability as a task-specific agent — an early signal for the product's reputation and adoption; but Mollick explicitly noted his usage was too limited for a detailed review, so early verdicts still await broader validation.

• OpenAI's Dot hands-on: proactive always-on agent reads Slack, flags conflicts, alternately flashy and frustrating — danshipper · 2026-09-30
• User review: dots delivers the best AI collaboration experience yet — BorisMPower · 2026-09-30
• Early Tester Says Specialized Dots Performed Surprisingly Well as OpenAI Expands Access — BorisMPower · 2026-09-30
• Dots review: OpenAI's always-on agents read your Slack and flag conflicts first — every · 2026-09-30
• Day-zero review: ChatGPT's always-on agent Dot reads your Slack, warns you mid-flight booking — every · 2026-09-30
• Hands-on with ChatGPT's always-on agent Dot: it read Slack to flag a flight conflict — every · 2026-09-30
• OpenAI dots hands-on: local tools, cloud Codex threads, one synced chat history everywhere — BenBajarin · 2026-09-30
• Matt Shumer on OpenAI's Dots: best-in-class AI, but UX isn't there yet — mattshumer_ · 2026-09-30
• Matt Shumer tests OpenAI Dots: best-in-class AI, UX still needs work — mattshumer_ · 2026-09-30
• Ethan Mollick on Dots: a capable model with your data is remarkably useful — emollick · 2026-09-30
• Ethan Mollick: OpenAI's Dots is 'remarkably useful' and more Clawlikes are coming — emollick · 2026-09-30
• OpenAI dots hands-on: one bot with all your context, cloud 24/7 — paw_lean · 2026-09-30
• Dots first impressions: a proactive personal-assistant agent with a transparent scratchpad — soumitrashukla9 · 2026-09-30
• OpenAI employee on dots: it noticed his canceled ride, nudged him to eat, and spawned Codex threads to scan PRs — gabrielchua · 2026-09-30
• OpenAI Dots First Impressions: High Latency and Rough UX Fail the Mum Test — cedric_chee · 2026-09-30
• Developer Tries OpenAI Dots After Grok Bot and Muse: Undercooked and Disappointing — cedric_chee · 2026-09-30
• OpenAI's always-on agent 'dots' day-one review: useful but buggy, free on Pro ($100+/mo) — johnseach · 2026-09-30
• Early Dots user review: 'an unpaid intern who's smart but doesn't care' about the job — DudeBroJustin · 2026-09-30
• People are overreacting to DevDay; Dots plus persistent agents are step one toward AI employees — haider1 · 2026-09-30
• Early power user on OpenAI's dot: always-on proactive AI companion unlocks real new value — iruletheworldmo · 2026-10-01

2 more related posts →

Episode 2 · OpenAI's New Dots Can't Even Read Public ChatGPT Links (2026-10-01, 3 posts)

Artist Sterling Crispin found that OpenAI's newly launched Dots cannot access ChatGPT conversations, even publicly shared links, drawing criticism on the product's day-one capabilities.

• Dot can't access users' ChatGPT chats, user finds on launch day — sterlingcrispin · 2026-10-01
• OpenAI's Dots Can't Read ChatGPT Conversations, Even Public Links — sterlingcrispin · 2026-10-01
• OpenAI's new Dots can't even read public ChatGPT conversation links — sterlingcrispin · 2026-10-01