Sponsored
vercat PokerGod Token Reset 3 Things Today
$5/week ↗
Projects

I Sent My AI Agent Fred a Photo. It Found the Exact Aisle

@justinmcclung

added

Just a quick example of a real world use case for AI assistants like ChatGPT Dots. I needed an inexpensive light socket, so I sent a picture to Fred, my dot in ChatGPT, and asked if Home Depot in Hays had something similar available that night. After one follow-up question, Fred found an option for $4.97, checked local availability, and gave me the exact aisle and bay. This short video recreates our conversation using the original product images. A simple example of how a dot can help with an everyday task. Hello Fred.

Justin McClung's Dot ("Fred" in ChatGPT) found a $4.97 light socket at Home Depot in Hays from a single product photo, checked local availability, and gave the exact aisle and bay after one follow-up question.

View on YouTube

More Dots use cases in Projects

See all 333 →
Dot scanned my Slack unprompted and offered a dev-handoff checklist

Dot scanned my Slack unprompted and offered a dev-handoff checklist

Dots를 오늘 처음 써 봤다. 이름 지어 달라고 하길래 이름 짓고 있는데, 그 사이에 나에게 도움이 될 만한 일을 찾겠다고 하면서, 나의 Slack 메시지를 뒤졌는지, "누구랑 어떤 작업이 예정되어 있는데, Slack에 흩어진 결정 사항을 모아서 개발 전달용 체크리스트로 정리해 드릴까요?" 이러는거다. Dots가 있어서 클릭하고 들어가기만 했는데, 내 정보가 순식간에 털리는 느낌이다.
@happytlog
Dot passed a stress-tested hiring coordination task

Dot passed a stress-tested hiring coordination task

I Gave an OpenAI Dot a Hiring Task, Then Tried to Trip It Up I handed an OpenAI dot a hiring task, changed the rules, and added a duplicate. Dots are OpenAI's always-on agents. Each has its own cloud computer and can coordinate work through ChatGPT Work or Codex. The test (all fictional): 4 application summaries, a role brief, and clear boundaries: organize evidence, keep gaps visible, leave decisions to the human. The catch it caught: applicant two claims Python, but the project was a browser interface with no Python sample. The dot kept both facts visible and marked the evidence unknown, without treating a claim as proof. Changed priority to project-first: it reordered the columns but changed nobody's eligibility. Added 2 more plus a duplicate: result was 6 unique applicants, not 7. Evidence stayed attached correctly. Asked for an interview draft for applicant one: 5 project-specific questions, invitation left as a draft with placeholders. We chose the applicant, not the dot. The honest part: one staged session, fictional data, no comparison to regular chat. It organized. The decisions stay human.
@opencvuniversity