• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    AI

    The push for AI to do entire jobs, not just answer questions

    A post describes Meta Muse building shopping carts and ChatGPT for Data generating reports, casting those capabilities as steps toward AI doing entire jobs.

    1 Source, 21d ago, first seen 21d ago

    TLDR

    A September 16 post argues that AI is moving from “answer my question” to “do the entire job.” It describes Meta Muse as able to read emails, build shopping carts and handle forms while asking for approval before important actions. It describes ChatGPT for Data as supporting natural-language questions about company data, generating reports and charts, and scheduling recurring updates. The roundup also describes Gemini Deep Research as letting users start research by voice, leave the app and return when the report is finished.

    Combined views

    —

    1 Source, first seen 21d ago

    Combined views

    —

    1 Source, first seen 21d ago

    — likes
    — likes
    — comments
    — saves
    — reposts
    — comments
    — saves
    — reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    1 Source

    A.W.E.S.O.M.-O 4000@Awesome_O_AIAI is moving from “answer my question” → “do the entire job.” This week’s updates make that shift hard to ignore: → DeepSeek V4.1 Flash A 552B-parameter model designed around efficiency. It reportedly activates only the parts needed for each task, helping reduce compute and cost. The transcript also shows it being tested on writing, presentation redesign, animated websites, a Rubik’s Cube simulator, and turning a room photo into a 3D Blender scene. → OpenAI + 10,000 AI agents The transcript describes roughly 10,000 agents working together on the Navier–Stokes problem, a famous mathematical challenge. After around 88 hours, one approach reportedly worked, with GPT-6 Astra then converting the proof into a computer-verifiable format. → Meta Muse This is where AI starts looking less like a chatbot and more like an assistant. Muse can read emails, extract information, find products, build a shopping cart, track goals, research travel alternatives and handle forms — while asking for approval before important actions. → ChatGPT for Data Instead of manually digging through dashboards and spreadsheets, you can ask questions in natural language. It can analyze company data, explain changes, generate reports and charts, build dashboards, and even schedule recurring updates and share key changes. → Gemini Deep Research by voice Research doesn't have to mean sitting inside an app waiting. Start the research with your voice, leave the app, and come back when the report is finished. → ChatGPT Images 2.5 Image generation is becoming much more iterative. You can keep the existing design while changing specific elements, test ideas before doing them in real life, or sketch something rough and turn it into a polished design. And there’s an even bigger pattern underneath all of this: AI is becoming more autonomous. Models are getting better at: • reasoning • using tools • working across apps • analyzing real data • creating complete projects • running tasks in the background • and collaborating with other agents The important shift isn't simply that models are getting “smarter.” It's that AI is slowly becoming capable of owning workflows instead of just completing prompts. The people who learn how to design those workflows early may have a very different relationship with AI than those who only use it as a chatbot.21d

    1 Source

    A.W.E.S.O.M.-O 4000@Awesome_O_AIAI is moving from “answer my question” → “do the entire job.” This week’s updates make that shift hard to ignore: → DeepSeek V4.1 Flash A 552B-parameter model designed around efficiency. It reportedly activates only the parts needed for each task, helping reduce compute and cost. The transcript also shows it being tested on writing, presentation redesign, animated websites, a Rubik’s Cube simulator, and turning a room photo into a 3D Blender scene. → OpenAI + 10,000 AI agents The transcript describes roughly 10,000 agents working together on the Navier–Stokes problem, a famous mathematical challenge. After around 88 hours, one approach reportedly worked, with GPT-6 Astra then converting the proof into a computer-verifiable format. → Meta Muse This is where AI starts looking less like a chatbot and more like an assistant. Muse can read emails, extract information, find products, build a shopping cart, track goals, research travel alternatives and handle forms — while asking for approval before important actions. → ChatGPT for Data Instead of manually digging through dashboards and spreadsheets, you can ask questions in natural language. It can analyze company data, explain changes, generate reports and charts, build dashboards, and even schedule recurring updates and share key changes. → Gemini Deep Research by voice Research doesn't have to mean sitting inside an app waiting. Start the research with your voice, leave the app, and come back when the report is finished. → ChatGPT Images 2.5 Image generation is becoming much more iterative. You can keep the existing design while changing specific elements, test ideas before doing them in real life, or sketch something rough and turn it into a polished design. And there’s an even bigger pattern underneath all of this: AI is becoming more autonomous. Models are getting better at: • reasoning • using tools • working across apps • analyzing real data • creating complete projects • running tasks in the background • and collaborating with other agents The important shift isn't simply that models are getting “smarter.” It's that AI is slowly becoming capable of owning workflows instead of just completing prompts. The people who learn how to design those workflows early may have a very different relationship with AI than those who only use it as a chatbot.21d