AI Copilots Everywhere; GPT's Trillion Dollar Economy

General Discussion
1

If you missed Microsoft’s Office 365 Copilot event this week, you might want to carve out an hour to read and watch the video. Be sure to watch it end-to-end with relatively focused attention because what you’re about to see is similar to the moment when the first web browser, Marc Andreessen’s 1993 Mosaic (later Netscape) was unveiled.

The fuse for the trillion-dollar economy was lit last year with fairly useful GPT examples. ChatGPT4 is simply an affirmation that this new economy is both significant and near. Microsoft simply validated - at an enterprise level - just how powerful LLMs can change everything.

Today, Microsoft is to LLMs what Netscape was to the world wide web.

While I have no strong affinity for Microsoft or their products since personally abandoning Windows in 2003, the vision they paint (The Future of Work with AI) is quite remarkable. It demonstrates that the rumours were fact; Microsoft has been quietly working with OpenAI and GPT models since early 2020 to be able to exhibit this much forethought and vision of future work with paired-AI copilots.

In 60 minutes, Microsoft wholly dismembered Google Workspaces, which now looks as if it were built by a team of five-year-olds who missed the Mensa cut. Heads will likely roll at Google, and they’ll catch up over time. Still, in the meantime, the tens of millions of businesses and hundreds of millions of users who jettisoned Microsoft Office to feast on a low-cost collaborative alternative to the office suite will be grazing in a pasture that is arid, brown, and unable to nourish the appetite we crave for advanced productivity in the AI economy.

Copilots Everywhere

But more important, micro pilots as well. The nature of AI and the implementation approaches vary. They can start small and grow in complexity. We’ve already seen some interesting experiments in the Airtable community, and the diverse possibilities to enhance work performance are almost infinite.

Copilots will soon exist everywhere, and if you aren’t using them or building them, you’ll miss the opportunity to participate on the ground floor of an emerging trillion-dollar economy. Worse, you’ll watch from the sidelines as your competitors make more deals, provide more advanced solutions, and outsell you on every level.

Google Workspaces and LLMs

I’ve been a huge fan of Workspaces since 2010. I have a historical client base of businesses that have created advanced automation solutions using Firebase, Google Apps Script, and other Google Cloud features and SDKs.

Google struggles to get ahead of the AI movement surrounding OpenAI and its broad array of GPT APIs. But, until it can offer something tangible from its own stable of AI research experiments, GPT will be an attractive and highly useful platform for building many of the examples Microsoft demonstrated this week.

Google’s rich development environment has allowed me to build impressive GPT features intersecting all Workspace document types.

  • Given a slide deck → write talking points for the presentation.
  • Given AI-generated talking points, create and insert an appropriate image for each slide.
  • Given a large collection of documents in Google Drive, build a search bot capable of locating and summarizing documents based on a simple natural language query.
  • With a corpus of Gmail messages, categorize and report metrics about the conversations.
  • With a new spreadsheet, use the column headings to generate 100 sample rows of data into the sheet.

The possibilities for integrating GPT features into Google Workspace workflows and documents are limited only to your imagination.

The Reach of Copilots

Some Airtable consultants eat their own dog food; they run their entire business on Airtable and use the same solution patterns they advocate for their clients. Most, however, don’t.

With AI and LLMs, it will be a very different climate; purveyors of AI solutions must use these future work patterns or risk disruption. Imagine a web design firm competing for business in 1998 without a modern-looking website. Imagine if Expedia executives used travel agents while promoting the disruption of the travel industry. Imagine if Apple execs stayed with Blackberry post-2007.

The depth and impact of AI are tantamount to deep-rooted disruptors we have witnessed in our lifetimes. This is not a superficial advancement that provides a new sheen for dull and dated application surfaces.

I’m old; my career should have ended a decade ago. Delightfully though, integration and automation demands have kept me somewhat relevant. Oddly, my near-fifty-year experience appears to be the prologue to a new, emerging, trillion-dollar adventure.

Like Microsoft, I quietly started learning everything I could about GPT, LLMs, embedding architectures, vector databases, and search index architectures four years ago. I sense I’ll be working with copilots to build copilots until at least 2036.

3 Likes
2

Your analogies never fail to make me chuckle :joy:

3 Likes
3

I wanted this one to suggest the IQ equivalent of a craps roll of loaded dice at Binion’s Casino. However, it’s not easy to convey quickly, and the rant would soon mimic a Dennis Millerism (e.g., v. to use too many large, complex words in a sentence). For example…

U.S. foreign policy makes about as much sense as Robert Fulton having sex with Beowulf at the First Battle of Antietam. I mean, when a neo-conservative defenestrates, it’s like Rovslofski filibustered deoxymonohydroxinate.

1 Like
4

Just checking to see if I watched the wrong video. The video I saw was 36 minutes long.

I think I better understand what you mean by copilot now. I like the term, especially the comparison to autopilot. I also found the slides showing how Copilot ties together Microsoft 365 apps, Microsoft Graph, and the Large Language Model to be educational. Copilots will be (are?) great tools. They are a major technological step into the future. History is full of technological advances that made it easier for more people to do more, better, faster, and cheaper. This is one of those things.

Here are a few thoughts that flitted through my brain as I watched the video.

  • I once read a dystopian novel where people didn’t know how to write or do simple arithmetic. The protagonist could read and appreciate poetry, but could not hand write the alphabet. Engineers could not do simple arithmetic without a calculator. It felt far-fetched when I read it, but seems less far fetched now. Having the general population loose skills as those skills become archaic is natural, necessary, and sad.

  • Computer programming has come a very, very long way. Early computer programs bore little resemblance to human language, but computer languages evolved to resemble English more and more. It makes sense to get to the point where people can write programs in English. I wonder what will happen to other languages.

  • As it gets easier to write and run a program, the more programs there are, and it eventually becomes cost effective to write complex programs that are run only once. (I really like Make’s “run once” button.) The system I saw is like writing a program (in natural language) that is compiled by the copilot and references lots of libraries (including personal libraries of data in Microsoft Graph), run on the LLM, and formatted for presentation by the copilot. I’m probably getting this all wrong, but that’s what flitted through my brain as I tried to make sense of what I saw.

  • I find it sweet that “Proseware” and “Northwind” are companies in the presentation. Like little Easter eggs or cameos.

  • Most people use only 10% of what PowerPoint can do? Really, they actually use that much? That seems really high. Decades ago, I used many features in MS Word that most people didn’t even know existed, and I still probably used less than 10% of what Word could do. On the other hand, I don’t see a problem with using only 10% of a programs features.

  • In my previous job at a software company, I wrote a lot of documentation that subject matter experts were supposed to review. I’m pretty sure that most of SME’s never actually read most of my documents. There was just too much text. There were also lots of documents with requirements and plans that most people didn’t read. With Copilot as presented in the video, people could generate mountains of text that will not be read, and people will use Copilot to summarize that text, and people will probably not read Copilot’s summaries either.

  • Physical tasks will be relegated to hobbies or elite athletes/performers. When the machines can do a physical task better, faster, and cheaper than people, why bother having people do the task at all? Answer: do the physical task if the physical task is fun (hobby), do the physical task if it can be monetized as entertainment (elite athletes/performers), or do the physical task if it is necessary for the upkeep of the human body.

  • While tools like copilot will make work easier in the short run, they will make it harder to stand out in the long term. If most people have access to tools that improve productivity, it is a rat race (Secret of NIMH reference) just to keep your position relative to everyone else. Even so, we should still embrace responsible use of better technology.

  • Writing papers is hard. Editing papers is hard. LLM can make both processes easier, and can make learning how to do both processes easier. But how many people will just turn all the work over to the LLM and not bother learning how to edit? I’m really curious what education will look like for future generations. I wonder if there will be debates similar to what happened with “new math” and “whole language” versus phonics. I am both hopeful and fearful of how AI can and will transform education.

  • I’m pretty sure that I won’t get access to Microsoft 365 copilot anytime soon. But I’ll keep my Microsoft 365 subscription that I was thinking of canceling.

2 Likes
5

The broadcast of the live event/presser was 60 minutes.

6

Not sure I follow this logic. And to be clear - tools like copilot - which “copilot” specifically. There are now about a hundred copilots.

Are you saying there’s a downside to a world with many diverse and deeply focused copilots?

One way a company can “stand out” is by posting profits in excess of their competitors. Ultimately, productivity, accelerated work, and automation is the measure of utilizing AI.

7

Maybe the presentation was cut down for the video?

8

I’m think more at the individual employee level. When everyone can create beautiful slide shows by typing a single sentence, making a slide show that stands out and is better than the rest will be harder to do.

This isn’t a criticism of better technology, just a general observation. We should still adopt the better technology because otherwise we fall even behind. But the technology won’t get us as ahead of our peers and one might think because our peers will be advancing too. And it will be harder for individuals to stand out because it will take doing non-trivial things to stand out when all the trivial things are automated.

9

Yeah, one could anticipate a homogenization of the workforce. But this was a sentiment I heard in the 1980s when WordPerfect was displacing Wang legal systems or in the 1990s when Windows was displacing DOS. Templates and wizards, it was believed, would make all output look pretty much undifferentiated. Everyone was wrong about this, of course.

And by “non-trivial” you mean actually composing a prompt that returns better information than your peers are able to compose. :wink:

This is precisely the nature of this post.

Despite my requests and often very stern admonitions for lax attention to written details and communique, all I get are excuses and sometimes the equivalent of cave dweller grunts. The decline in the use of words and well-crafted compositions has been occurring over many decades. Now we have pseudo-tokens for language; the affable emoji without context or background are generously used in place of thoughtful conversation.

While this decline in writing among younger generations is slow and almost imperceptible, what comes next will be abrupt.

To leverage LLMs, you need to write well because to make it work to your advantage, it needs to understand what you want. If you can’t express clear thoughts and ideas with words and some punctuation, your experience with the newest and likely the most profound technical innovation in decades will not serve you well. Instead, it will serve you in a mediocre way, and you will struggle to stand out in your field.

I believe exactly the opposite will occur - LLMs and copilots will help truly remarkable workers do their best work and lots more of it. The gap between mediocre and hyper-productive workers will widen. Over the next decade, it will be very easy to spot the output of workers who have wholly embraced LLMs, and this bubble will be comprised of the strongest communicators.

10

I also believe this. But I also believe that how educators leverage LLMs to teach will make a big difference. Students of teachers who figure out how to effectively leverage LLMs will have a huge advantage over those whose teachers cannot.

1 Like
11

Astonishing. I watched the video and I wasn’t ready for that.

We all dreamed about such a system that could help us, but I didn’t expect something so fine-tune to see the day so quick, it’s the first time that Windows actually surprises me like that in a good way.

And yeah, humans will run on copilots, that much seems inevitable. But what will actually happen with that, nobody can foresee.

I also dream of a personal copilot, that would help me in my daily life, and this seems closer than ever.
And I bet there will be different such “Copilot buddies”, with different capabilities, and pricing. And, once again, the richer will get the best, and the gap will widen even more with the rest of the population.

This is a probable future, too. And it could be as close as 5-10 years from now. And yet I’m being pessimistic about it.

12

We’ve been talking about AI for a few decades now, and quietly, we are making slow progress. Nothing really big, nothing really impressive. Last year, we hit the"knee" of the curve.

CleanShot 2023-03-20 at 18.41.15@2x
CleanShot 2023-03-20 at 18.41.15@2x1920×1922 178 KB

Now advancements are going to speed up.

2 Likes
13

More thoughts….

Getting good results from an LLM requires prompts that include

  • quality data inputs, such as the source files that were fed to the copilot
  • clearly defined outputs, such as the format and length of the output defined in the prompt
  • editing skills to determine which parts of the output to keep, which to have the copilot revise, and which to revise without the copilot.

Good development has always needed these three things. In education, figuring out how to do something is often the focus of teaching. Copilots and LLMs help shift the focus from how to do a thing to what inputs do you need and what is the desired output. That sounds a lot like a design doc without a lot of the tedious details.

14

… and in some cases to actually doing that thing.

15

And yet… What Microsoft has showcased is even more powerful than that, because they “improve” on-the-fly the initial prompt, to enhance it, probably fixing typos, reworking it in such a way that the intent of the user is properly conveyed.

So, being an expert in such a prompt might not even be needed.

16

Microsoft is well aware that prompt engineering should generally not be used by end users. It is brittle, unpredictable, and will not likely exist in GPT-6 or 7.

The more they can insulate users from this voodoo, the better their services will be.

1 Like