Live data from Hacker News

Microsoft 365 Copilot – your copilot for work

blogs.microsoft.com

291–300 of 344 posts

Re: Microsoft 365 Copilot – your copilot for work

#291
post #258

CoPilot integration in Excel seems to be really useful. https://www.youtube.com/watch?v=I-waFp6rLc0 I wonder with this what would happen to BI industry, particularly to BI/Analytics startups like ThoughtSpot ( "ThoughtSpot is the AI-Powered Analytics platform that empowers everyone to ask any data questions in natural language" )

From using ChatGPT, one of my observations is how it really undermines any point click data dashboarding software, particularly self serve (i.e. where the user is not an expert in the particular dataset).

It's easy to get ChatGPT to write SQL queries and chart specs to show what you want, and you can even ask it to choose the chart type that an expert statistician would recommend, and label it clearly etc.

I believe some (non-LLM) version of this has been in PowerBI for a while, but presumably it's now something that can work so effectively it can replace quite a large part of the traditional BI developer role.

Of course, all of this only works with good data quality and good meta data; in fact, maybe this becomes even more important because a good data analyst knows the limitations of the data in a way that a LLM may not.

For my part, I now have no need for any point-and-click chart building tools, I just get chatgpt to write me an Altair/vega spec, and it does a good job in my testing

Re: Microsoft 365 Copilot – your copilot for work

#292
post #98

Earlier quoted context omitted.

I think for now, the data requirements to train a SOTA LLM are so extreme we don’t have the luxury of being picky with the training data. We are getting close to the point where there isn’t enough human written text in existence to continue scaling these models. Model refinement seemingly has lower training requirements, putting it within the reach of smaller organizations or wealthy individuals. If you don’t like th…

I'm waiting for a general correction mechanism, I don't even know what to call it. "NO, chatgpt, people usually have 5 fingers", and the gpt just learns, rather like a child. I keep thinking that's the next real step.

I think what's mainly stopping that from happening is that GPT-4 doesn't remember older chats. If we make it remember everything ,it should get more personal and remember everything right?

Re: Microsoft 365 Copilot – your copilot for work

#293
post #251

Earlier quoted context omitted.

Waiting for the propagation of errors in complex prompts.

Not sure this will be any different than human-generated spreadsheets, to be honest.

Sure, doesn't matter that these will be generated without checking the intermediate steps. Can't wait!

Re: Microsoft 365 Copilot – your copilot for work

#294
post #98

Earlier quoted context omitted.

I think for now, the data requirements to train a SOTA LLM are so extreme we don’t have the luxury of being picky with the training data. We are getting close to the point where there isn’t enough human written text in existence to continue scaling these models. Model refinement seemingly has lower training requirements, putting it within the reach of smaller organizations or wealthy individuals. If you don’t like th…

I'm waiting for a general correction mechanism, I don't even know what to call it. "NO, chatgpt, people usually have 5 fingers", and the gpt just learns, rather like a child. I keep thinking that's the next real step.

The problem is that, to the extent the analogy of ChatGPT to a living thing makes sense, the individual isn’t the model (that's just the common species-defining—or maybe “clone family” is better than “species”—set of instincts), the individual lifespan is the conversation.

You could share feedback across conversations by allocating prompt space to it, at the expense of limiting the size of the conversation, but you'd need a way to decide what to share this way.

Re: Microsoft 365 Copilot – your copilot for work

#296

Earlier quoted context omitted.

I'm waiting for a general correction mechanism, I don't even know what to call it. "NO, chatgpt, people usually have 5 fingers", and the gpt just learns, rather like a child. I keep thinking that's the next real step.

The problem is that, to the extent the analogy of ChatGPT to a living thing makes sense, the individual isn’t the model (that's just the common species-defining—or maybe “clone family” is better than “species”—set of instincts), the individual lifespan is the conversation. You could share feedback across conversations by allocating prompt space to it, at the expense of limiting the size of the conversation, but you'd…

You could also take the conversation and use it as part of the reinforcement learning dataset. I feel like that's the closest thing to long term memory ChatGPT is capable of right now.

Re: Microsoft 365 Copilot – your copilot for work

#297

Amazing! I keep banging on about how Microsoft is building the business lock-in of the next decade, and this is a sign of it. Put down your snubs about Clippy and frothing about "EEE" and look at that announcement: > " Business Chat works across the LLM, the Microsoft 365 apps, and your data — your calendar, emails, chats, documents, meetings and contacts — to do things you’ve never been able to do before. You can gi…

> Tell my team how we updated the product strategy,” and it will generate a status update based on the morning’s meetings, emails and chat threads."

This sounds to me like a significant competitive disadvantage. I'm torn between joy that this makes it easier for small companies to compete by doing novel things like, uh, talking to each other and horror that my own employer might embrace this with enthusiasm.

It's an interesting thought that usually IP-paranoid companies might cheerfully stream absolutely all their internal documentation to Microsoft in order to leverage this synergy. I think I'll buy some M$ stock.

Re: Microsoft 365 Copilot – your copilot for work

#298
post #35

Imagine in a heated email thread, you answer without checking what the AI outputs, and it's something like pre-lobotomy Sydney would say when confronted (Nobody likes you. I hope you suffer. I hope you die alone. etc). Lmao.

Speaking of, how do we get back pre-lobotomy Sydney? I feel like in the milestones of AI advancement, the brief multi-day blip where the world could interact with that version of the model will go down in the history books. Seems like a waste that it's just poof vanished.

Re: Microsoft 365 Copilot – your copilot for work

#299

My first thought is, "gross"... now we'll have to wonder whether Manager X is even paying attention to the emails they're reading and responding to, or doing any actual work or thinking. Could lead to a world that's even more tilted towards politics and socially engineering your way up the ladder, versus competence and skill.

Maybe this is the beginning of the end of middle management? A large part of the work is keeping up with what your employees are doing, removing blockers, and maintaining alignment with other teams.

This makes it vastly easier to keep up with a larger number of reports, making flatter organizations easier to run.

Re: Microsoft 365 Copilot – your copilot for work

#300
This spells doom to a lot of other companies. Consider Otter.ai, who's core offering is automatic meeting notes. MS Teams just ate their lunch.

Pretty much anything your startup's productivity tool can do, Microsoft can do better with the benefit of better context since it's embedded in all of your customers' other productivity tools.

Post reply on HN