Microsoft to demo its new ChatGPT-like AI in Word, PowerPoint, and Outlook soon

[highplainsdem]

Source: The Verge

Microsoft is getting ready to demonstrate how its new ChatGPT-like AI will transform its Office productivity apps. After announcing and demonstrating its Prometheus Model in its new Bing search engine earlier this week, Microsoft is gearing up to show how it will expand to its core productivity apps like Word, PowerPoint, and Outlook.

-snip-

Microsoft is moving quickly with this integration mainly because of Google. Sources tell The Verge that Microsoft was originally planning to launch its new Bing AI in late February, but pushed the date forward to this week just as Google was preparing its own announcements. Google then announced its ChatGPT rival Bard a day ahead of Microsoft’s event.

-snip-

Nadella’s push for AI across Microsoft’s products is driven by the consumer response to ChatGPT. Analysts at UBS estimate that ChatGPT reached 100 million monthly active users after just two months. More than 1 million people have signed up for the Bing waitlist in 48 hours, and Bing was the third most popular app in the App Store in the US as of Thursday.

Microsoft wants to seize the AI opportunity throughout 2023, and unlock some of the features and power of its Prometheus Model that it hasn’t publicly discussed yet. Expect AI to be the big buzzword for the entire tech industry throughout the rest of the year, as Microsoft and Google kick off a battle for the future of search and AI.

Read more: https://www.theverge.com/2023/2/10/23593980/microsoft-bing-chatgpt-ai-teams-outlook-integration

---

This crazy AI-race is heating up fast despite the mistakes AI is making - not to mention the disruptions it will cause.

From an article in The Verge two days ago, about the AI included in the new Bing and Edge having the ability to write things for you:

https://www.theverge.com/2023/2/8/23590873/microsoft-new-bing-chatgpt-ai-hands-on

There’s even a compose tab in this new sidebar that makes all the work of typing in parameters a little easier, offering quick options for users to specify tone, format, and length (rather than having to type these qualities out by hand, as you do in ChatGPT). You can ask Microsoft’s AI model to write about anything here, and it will be neatly formatted and ready for a blog post, email, or a simple list.

So I guess Microsoft wants us to all get used to email and website exchanges (here on DU, too, probably) being written not by humans, but by AI that will also helpfully suggest tone, format and length.

Imagine entire discussions being mostly argued by AI.

That article from two days ago included what the journalist called "an impressive and genuinely funny" resignation letter he had asked Microsoft Edge's new compose feature to write. He'd requested the funny tone.

From that letter:

So, I’ve decided to save you the trouble of firing me and just quit. I don’t want to compete with a machine that can crunch numbers, generate reports, and make witty jokes at the same time. I don’t want to work for a boss who values AI over human intelligence and creativity. I don’t want to be part of a company that is willing to sacrifice its employees for the sake of efficiency and profit.

I’m sure you’ll be very happy with your new AI assistant. I hope it doesn’t turn against you and take over the world. I hope it doesn’t make any mistakes that cost you your reputation and customers. I hope it doesn’t get bored and start playing games with your data. I hope it doesn’t develop a personality and fall in love with you.

But if it does, don’t come crying to me. I’ll be too busy enjoying my life, pursuing my passions, and finding a new job that appreciates me for who I am.

I don't find that terribly funny. Especially given the happy ending of how wonderful life will be for replaced workers as AI takes over.

I don't think the reporter who wrote that will be laughing if he does lose his job to AI.

And I'm not looking forward to a world where what you'll see online from people you encounter will be less and less likely to be their own writing, and more and more AI-generated responses designed to make them seem different than they are.

I'm okay with my own writing. I will not be using AI to substitute for it.

I hope others here won't be using AI to write their DU posts.

But you can expect to see it more often. And one sad result is that even the stupidest MAGAts will be able to use it to make their tweets and comments on other websites seem more intelligent and literate. And they'll be able to post more of them.

[mahatmakanejeeves]

1. Bing AI sounds like a circle of hell even farther in than Dante imagined. NT

[Renew Deal]

7. Why?

The search results with chat are optional.

[Renew Deal]

8. I've seen recent analysis on this topic that said that people will not lose jobs because of AI

What will happen is that people that know how to use AI will replace those that don't.

[Yavin4]

20. Did ATMs put bank tellers out of work?

ATMs have been around for decades, and there are more retail banks and tellers than ever.

[Renew Deal]

23. Some of them

Though it's not a great comparison. Direct Deposit also put a lot of them into retirement.

AI will apply to almost all roles on some level. Bank tellers apply to banks.

[Yavin4]

25. There are more retail branches and tellers than before.

And most of them even have weekend hours now. Every tech transformation has created more jobs than what existed before. Often, they've created jobs that never existed previously.

Direct Deposit created new roles in accounting, cybersecurity, and website design for banks.

[Renew Deal]

26. Right and those that don't keep up with the skills

will fall behind.

[tinrobot]

16. It doesn't think for you any more than a book "thinks" for you.

It is just another way to gather information. You get to choose whether that information is useful or not.

[Karma13612]

29. But if you don't know

The information is in accurate, then you are relying on misinformation.

That’s not good.

[tinrobot]

44. There was plenty of misinformation before ChatGPT

It's not a new thing.

And no, it's never good. Always double check your sources.

[BootinUp]

3. this is just another tech fad. nt

[Renew Deal]

10. People said the same thing about PC's and then the internet.

Those people were very wrong and you will be too. AI is already very much a part of our lives. What's different about this is that it can intelligently interact. We have barely scratched the surface. Some things to come: Better home assistants/companions, AI business analysts that are better than most human analysts, better automated customer service than what we currently have, and lets not get into the self-driving car stuff. This is around 1993 in the internet age. The next 5 years will be amazing.

[BootinUp]

12. Not a very good example (gibberish AI generated)

Last edited Fri Feb 10, 2023, 01:40 PM - Edit history (1)

Hey I is a black box we don’t know what it’s going to produce when it will produce something valuable and when it will produce something that valuable this is a completely different than a an object like a personal computer.

[Yavin4]

21. People said the same thing about the internet in the mid 1990s.

Ten years later, an entire new class of billionaires was born based primarily on their ability to adapt the power of the internet.

[BootinUp]

22. There is something I am not communicating obviously.

The internet and a personal computer are tangible things with clear specifications. The other is a black box of voodoo to everyone but the designer. I cannot make it clearer. There will not in our lifetimes be AI that is used for any kind of critical work that it is not already considered capable of.

[Karma13612]

28. But it generates

Inaccurate information. And if you don’t know the information is wrong, then you are relying on disinformation.

I see this as a boon for anyone with bad intentions.

[NBachers]

4. Imagine sentient Magats telling it to cruise chats looking for certain terms, and formulate comments

Automated Maggot Robots

[XorXor]

18. Yeah, even with what we have now, it really lowers the barrier of entry for creating disinformation

Just give it some prompts to ensure it covers certain topics with a general idea of how they want it to discuss them. It will output some text that does a competent job at saying what they want it to say, and every time they run that prompt they end up with something a little bit different. Which makes it harder to detect such activity. They could even include context of the current conversation to make it seem even more legit.

It's not hard to automate that either. There isn't even a requirement to think once the tool is built. One thing is that the higher-end GPT-3 models are not publicly available and the costs can add up if used on a large scale. Although, there are some legit open source models like GTP-J and GTP-Neo that are not nearly as large, but are still competent, that can be run on hardware that is in theory doable for regular people. Still needs a large amount of RAM and power GPUs with plenty of memory, but we're talking like 3 or 5 thousand dollars in hardware. I could see someone running that out of their garage, having the thing connect through tor to post from IPs all around the world.

[FakeNoose]

35. I think it has been happening already, but live people were doing it

Before we were calling them "MAGAts" there have been Chump-supporting operators (people or bots) posting the rightwing talking points and Chump-style insults on American chat forums. I can remember back to 2015 and 2016, before I came to DU, and for a brief time I was on the Yahoo discussion boards. This was during the time of Chump's campaign and most of us believed he had no chance of winning the Presidency. But I was amazed to see that the posters on Yahoo seemed to be overwhelmingly pro-Chump to the point of chasing away anyone with a liberal or Democratic viewpoint. (Including myself.)

However while I was there I noticed that many of the posted replies were quite similar to one another, like someone was copying and pasting text from an archive of rightwing insults somewhere. No matter what the topic under discussion, the Chump-supporting posters (all using different names, but sounding very similar) would chime in with the same insults and rightwing talking points. If someone such as myself, would challenge or question their statement, it was crickets. No answer. They had already moved on, and had no desire to engage with anyone.

I kept thinking this was such strange behavior, until we learned - after the election was over - that Russian agents had been acting on Chump's behalf on American SMS chat-boards for over a year. Suddenly it made sense to me. These Russians weren't "writing" anything because they couldn't. They were just copy-pasting stuff that some disaffected right-wingers had composed to insult the liberals with. The Russians had no desire to engage with anyone or display their ignorance of the English language. They were just posting the pre-written rightwing garbage everywhere they could, in as short a time as possible.

Amazingly, after Chump "won" the election in 2016 these rightwing posters - the ones that I suspected were Russian - disappeared almost immediately. By that time I had discovered DU and moved quickly off the Yahoo boards. So I can only guess that they were reactivated for the 2020 election. In 2016 it was probably live people doing the posting, but would be a short step to create an AI program to do the same thing. It would have been the Russians that took that first step, creating something to mess with the American SMS. Now the new Chat-GPT programs are quite a bit more sophisticated, but they had to start somewhere. These AI chatbots have been in the works for several years.

[dalton99a]

5. NEXT: Microsoft introduces AI readers for schools and businesses to parse AI-generated bullshit and

REALLY read the AI-generated material

[AllaN01Bear]

6. when it says ,,

,,,
i quit.

[AllaN01Bear]

9. since when did m$ buy bing search engine? and i think i will keep my simple word processor

. thanks

[Renew Deal]

11. The created it.

[Cheezoholic]

13. Oh no, the phreakin' pain in the ass paper clip is back

with "brainz"

[NotASurfer]

15. Clippy 2.0

Now with 20% more effective annoyance?

Can't wait for Clippy 3.0, with integrated T-800 Terminator functionality to put an end to those pesky end users and all their end user errors once and for all

[BadgerKid]

37. I, for one, welcome our new Clippy overlord

[Torchlight]

14. I wonder what the full consequences of this will turn out to be.

I can see some convenient uses for it in my household, but seeing that just about every tech advance exploited by bad actors, I'm also morbidly curious about what form and method the bad-faith and malicious users will use to accomplish their goals.

I never saw the internet coming, let alone its full impact on everything from international commerce to neighborhood friends, and I'm often taken by surprise not with the tech advance so much as the social uses it fills (or leeches into, dependent on perspective).

So I've learned not to predict anything that will come to pass in the digital world other than it will surprise the crap out me.

[Layzeebeaver]

17. So I just went and asked ChatGPT a question...

Me: tell me something interesting to post on the democratic underground forum


ChatGPT: Here's an interesting topic you could post about on the Democratic Underground forum:

"The Rise of Progressive Politics: How the Democratic Party is Evolving to Meet the Needs of 21st Century Voters"

In recent years, there has been a noticeable shift within the Democratic Party towards progressive policies and values. This has been driven in large part by younger voters and people of color who are advocating for bold action on issues such as climate change, economic inequality, and systemic racism. As a result, many Democratic politicians are embracing more progressive positions on these and other issues, reflecting the growing power and influence of these communities within the party.

You could discuss how this shift towards progressivism is changing the Democratic Party and the role that activists, voters, and political leaders are playing in this transformation. You could also explore the challenges and obstacles that progressive Democrats face in achieving their policy goals, and consider the future of the progressive movement within the larger Democratic Party.

Me: Why did you recommend this specific topic among any others?

ChatGPT: I recommended this topic because progressive politics has become an increasingly important and controversial topic within the Democratic Party in recent years. The rise of progressive politicians and movements like the Squad has generated a lot of debate and discussion about the direction of the party and what it stands for. Additionally, many of the policy issues that progressives are advocating for, such as climate change, economic inequality, and systemic racism, are among the most pressing and relevant issues facing the country today. By exploring the rise of progressive politics, you can engage in a meaningful and timely discussion about the current state of the Democratic Party and the future of American politics.

Me: now can you tell me all this in the style of a broadway musical?

ChatGPT: Of course! Here's the same information set to the tune of a Broadway musical:

(Verse 1)
The Democratic Party is changing its tune
With progressive voices rising in June
They're calling for action, on issues so dire
Climate change, inequality, and systemic fire

(Chorus)
The rise of progressive politics, you see
Is transforming the Democratic Party
With activists, voters, and leaders all in play
A bold new future, is here to stay!

(Verse 2)
The Squad is leading, with passion and might
They're fighting for change, with all of their might
They want a better world, for you and for me
And they won't stop until it's history

(Chorus)
The rise of progressive politics, you see
Is transforming the Democratic Party
With activists, voters, and leaders all in play
A bold new future, is here to stay!

(Bridge)
There may be challenges, on this road ahead
But progressives won't let their dreams be dead
For a brighter tomorrow, they'll fight to the end
And their message will ring, for all to attend

(Chorus)
The rise of progressive politics, you see
Is transforming the Democratic Party
With activists, voters, and leaders all in play
A bold new future, is here to stay!

[BootinUp]

24. How much of that is plagiarized?

[Layzeebeaver]

30. OK let's ask...

Me: in relations to the topic of topics for a democratic underground forum posting, can you list your sources and of any of the text your provided is plagiarised?

ChatGPT: As an AI language model, I do not have a personal memory or knowledge of the sources of the text data I was trained on. I have been trained on a large and diverse corpus of text data sourced from the internet, including websites, books, and other written material.

It is possible that some of the text generated by me could be similar to text found in other sources, including websites and articles, due to the large and diverse nature of my training data. However, this does not necessarily mean that the text is plagiarized.

OpenAI takes the issue of plagiarism very seriously and has taken steps to ensure that its language models are trained in a way that minimizes the risk of generating plagiarized text. For example, the model has been trained to not include direct quotes from specific sources, and the training data has been filtered to exclude content that is likely to be protected by copyright.

If you have concerns about the originality of the text generated by me, I encourage you to verify the information using additional sources and to consult with a subject matter expert.

[BootinUp]

34. Oh I get it vulcans and AI bots don't lie. lol

[Layzeebeaver]

39. flippant remarks that do not add to the logical discussion are useless

Do your own tests.

Register to access the chatbot.

Carry out your own questions.

But keep your passive aggressive sarcasm at bey until you have some real-world experience testing this ChatGPT tool.

All you are doing at this point is reinforcing conspiracy thinking.

[BootinUp]

40. You have things all upside down Layzee

Its a simulation. There are no proofs you can obtain that way.

[Layzeebeaver]

41. If paying attention to facts, and avoiding false conclusions is upside down, then I'm happy...

What 'simulation' are you talking about?

What "proofs" are you convinced that I attempting to obtain?

or (a better chance this has a bit of truth to it) are you just trolling me here to get a reaction that you can riff on?

Doesn't matter, hope you enjoy your day.

[NBachers]

19. Well, I guess I'll just program it to read everything for me while I go off and let my brain moulder

[Layzeebeaver]

42. There is actually a future possibility that touches on this theme...

There could come a day, when people can "direct" an AI tool to comb thru the internet on a daily basis and synthesise a compilation of news and information covering a specific topic.

So, if the topic was "The War in Ukraine", the tool would identify any and all internet listings (say for the past week) and present them all as a single summary.

Today, I can spend hours of my life each day searching sites. In this possible (not guaranteed) future I could have a daily program running to collect all internet info and present a summary to me every morning (with notations and links) - then I can choose based on my own criteria what I want to read and explore.

Today we rely on Google, bing, etc. and the programmers there to configure and run the search engines. In the future, they may become individually configurable, And yes, before anyone says it, there are a whole series of issues that come along with that....

Trusting your own self to come up with effective search criteria vs. believing that a bunch of AI engineers have your interest at heart is something to ponder...

[Karma13612]

27. I have no desire to use Chat GPT

Unless I just want to try for “entertainment purposes”.

Not allowing it to write correspondence for me and I will not use it when searching for information.

It’s quite possible I/we have come to the pinnacle of stupidity/evolution and are now on a downward trend.

I don’t think it’s my age either. I’m generally thrilled about innovation and scientific breakthroughs, but ChatGPT will be used incorrectly and will cause much harm.

Just like with cell phones and tasers, this is new technology without proper study and over site.
Consequences like driving while using cell phones. And tasers causing heart attacks and death, and abuse by LE.
The laws for cellphone use didn’t happen until bad behavior was well entrenched in our culture. And taser use is just a total mess.

[Layzeebeaver]

31. I think...

It could become an excellent source for inspiration if used correctly and in a positive manner.

It is not currently available as a search engine, as it has no access to the internet.

[Karma13612]

32. Then I must be misinformed.

I was under the impression it has been tested, or it’s functionality has been tested. People ask it questions, and it gives answers by searching the internet. And the results have been very very poor. The results are mostly full of wrong information.

https://www.pbs.org/newshour/science/analysis-chatgpt-is-great-at-what-its-designed-to-do-youre-just-using-it-wrong

[Layzeebeaver]

33. no. it was trained on huge text sources

but not the entire internet.

Also, it has zero access to any training data from beyond 2021.

Also, It cannot currently access the internet. so it cannot show you a web page.

It's still in "beta" mode, undergoing monitoring, etc.

But anyone can register and play around with it.

I used it to generate some disclaimer test for a slide deck on airport cyber-security. Given the current input, it did a rather good job with only a few changes on my part regarding scene setting and my personal style.

It's a cool tool. But that's it. It's a tool to be understood and used correctly.

[Karma13612]

38. Thanks for the clarification!

As with everything, it’s always important to know the weaknesses as well as the strengths of new technology!

[ruet]

36. The First, Real, Reason to Upgrade Office Since 2010.

I use Chat GPT as an advanced search engine. It aggregation and presentation of fact and statistic based information is really something else.

[Layzeebeaver]

43. ChatGPT has no access to the internet now

so be careful. all information since 2021 is not part of ChatGPT.