And feed stuff into it via an API. I always like to automate stuff. I guess you could use power automate but I'm not really big on the MS ecosystem (despite being part of a team that owns it at work)
Yes, totally different. If you use windows, there’s a pane in their Edge browser that opens up an AI chat window. You can use it for most of the things you get with paid OpenAI subscription, but it’s free for every windows user. They call it Copilot.
Most recent major feature release has it now (I don’t know what they call them anymore) but it showed up in my taskbar after an update maybe just over a month ago.
Yes. "Copilot" is just chatbot, now. The copilot that helps you code is now called "Github Copilot". I guess Microsoft couldn't leave a good name alone.
I believe Microsoft reused the Copilot brand that it acquired from GitHub for its AI assistant which is embedded in Windows, Bing, and apparently now Office. GitHub Copilot also still exists but is a different product.
> It makes it very difficult to navigate their offerings.
Microsoft has always done stuff like this. For instance, it was hard to know if you needed Windows Deluxe Professional Edition, Windows Master Programmer Developer Edition, or Windows Demigod Master of the Universe Edition (those might not be the exact names :-)
I suspect that the goal there was to confuse people enough that the wealthy bought the most expensive one just so they were sure they "got everything", while not losing the sale to the poor schlub who could only afford Windows Broke-ass Fool You Suck Edition.
There are a lot of “Copilot” extensions for VSCode, including Github Copilot, but none for Microsoft Copilot, AFAIK.
> It’s possibly confusing,
No, after Github Copilot being around for a while, Microsoft rebranding everything that is AI assistants “Copilot” (with “Microsoft” in some contexts, but often just bare) is confusing, especially since it follows lots of other projects adopting the X Copilot thing in between “Github Copilot” becoming popular and Microsoft introducing their new Copilot branding.
I suspect that it's got to do with excess availability of compute. Usage hasn't grown past the initial novelty wave and enough enthusiast programmers are recently talking about how Claude 3 performed better in their usage.
Get users hooked with free access. Wait few years. Start having lawyers tweak terms of service. Charge corporate customers a shit ton of money as free users sell their junk. Yank “free” tier, with none to minimal notice due to ToS. Cite limited resource constraint. Stick previously free users with $99/month access.
Or perhaps “chat” just isn’t the best use of LLMs. Autocomplete-based integrations, rather than chat, could be better. I know I love and get great utility from GitHub Copilot, but not really any online or offline LLM chatbot.
Also, from a UX perspective, I feel an autocomplete-based UI encourages me to make sure the thing “I’m saying” is actually what I mean to say - ie, to not implicitly trust LLM hallucinations.
LLM-based generative text everywhere, in every app, could be useful, very straightforwardly integrated at an OS level into all text controls, something MS is well-poised to do, and “Copilot” a very good brand for that. (And I would be shocked if Apple and Google aren’t working on the same.)
I use llms to write text for me sometimes simply to rewrite it myself because sometimes having something to criticise is better than having to create from scratch
Given the relative simplicity of integrating LLM autocomplete into OS text fields (compared to other integrations supported in modern OSes), I would be shocked if a state entity with balls (like the EU) doesn’t introduce laws enforcing that OSes must have open support for third-party LLM/generative text integrations.
That said, that’s almost certainly Microsoft’s short-term strategy until then! And probably part of why Apple made their sudden priority shift towards LLM features.
Given the ongoing issues with ChatGPT (slow and _incredibly_ lazy) and all the copilot nonsense, as well as my growing aversion of how MS are muscling in on this space, I’m about to make a wholesale switch to Claude for my API and chat access
47 comments
[ 2.8 ms ] story [ 65.5 ms ] threadjust either accept or reorient to selfhosting
https://fortune.com/2023/02/21/bing-microsoft-sydney-chatgpt...
Edit how lazy is the question, found the stand alone app
iOS: https://apps.apple.com/se/app/microsoft-copilot/id6472538445... Android: https://play.google.com/store/apps/details?id=com.microsoft....
It's possibly confusing, or just very simple. I'm not sure :-D
Microsoft didn’t acquire Copilot from Github, Microsoft introduced it at Github long after Github was part of Microsoft.
But I think it is kinda similar to how OpenAI is trying to call everything backed by their models a "GPT"
Microsoft has always done stuff like this. For instance, it was hard to know if you needed Windows Deluxe Professional Edition, Windows Master Programmer Developer Edition, or Windows Demigod Master of the Universe Edition (those might not be the exact names :-)
I suspect that the goal there was to confuse people enough that the wealthy bought the most expensive one just so they were sure they "got everything", while not losing the sale to the poor schlub who could only afford Windows Broke-ass Fool You Suck Edition.
This is about Microsoft Copilot – https://copilot.microsoft.com/ – not Github Copilot – https://github.com/features/copilot
There are a lot of “Copilot” extensions for VSCode, including Github Copilot, but none for Microsoft Copilot, AFAIK.
> It’s possibly confusing,
No, after Github Copilot being around for a while, Microsoft rebranding everything that is AI assistants “Copilot” (with “Microsoft” in some contexts, but often just bare) is confusing, especially since it follows lots of other projects adopting the X Copilot thing in between “Github Copilot” becoming popular and Microsoft introducing their new Copilot branding.
It’s gonna be ruthless between Google, Apple and Microsoft.
MSFT stock pumps hard for the next few years.
Like you said, they're trying to get people hooked.
But if AI isn't any useful to people in the first place, this goes nowhere and they're still stuck with a low/decreasing amount of users.
Also, from a UX perspective, I feel an autocomplete-based UI encourages me to make sure the thing “I’m saying” is actually what I mean to say - ie, to not implicitly trust LLM hallucinations.
LLM-based generative text everywhere, in every app, could be useful, very straightforwardly integrated at an OS level into all text controls, something MS is well-poised to do, and “Copilot” a very good brand for that. (And I would be shocked if Apple and Google aren’t working on the same.)
That said, that’s almost certainly Microsoft’s short-term strategy until then! And probably part of why Apple made their sudden priority shift towards LLM features.