The "Yes And" Slack Bot

I wrote a blog post a while back about collaborating on a list in Google Sheets. At USV, we do a lot of list making in Google Sheets. But Sheets doesn’t have a great conversational interface for coming up with new entries for the list. So we use email to do that but the process is clunky. In that post, I suggested that we might write a slack bot that takes a conversation in a slack channel and turns it into a list in Google Sheets.

A few days later, I got an email from Alex Godin who had built exactly that bot. We tried it out at USV, made a bunch of suggestions on how to make it better, and the result is the “Yes And” Slack Bot. It was approved yesterday by Slack and is now in their app store.

If you and/or your company uses Slack and Google Sheets a lot, you should give it a try. Details on how to do that are here.

I love it when a blog post at AVC turns into something. It happens a lot actually.

#mobile#Web/Tech

Comments (Archived):

  1. JimHirshfield

    Nice. I’ll add this to my list of bots to check out someday. I have a special spreadsheet just for that purpose.

      1. JimHirshfield

        I’ve had many an internet conversation with that robot.

      2. Lawrence Brass

        qwerty ytrewq?

      3. jason wright

        a brute force password attack? safe.

    1. Alex Godin

      Haha! Thanks Jim. Let me know if you have any questions when you try it out.

    2. ShanaC

      can I see the spreadsheet? it sounds useful

      1. JimHirshfield

        Are you as gullible as I am sarcastic?

        1. Anne Libby

          Or is she stealth sarcastic?

          1. ShanaC

            Unfortunately, rarely

        2. ShanaC

          Sometimes Yes?

          1. JimHirshfield

            Why the question mark at the end of your comment?

          2. ShanaC

            Slightly embarrassed that it’s true.The worst was when I was a freshman in high school in the school play and I fell for something one of the junior boys in said same play said 2x in a row in under an hour because I really wanted him to be honest and nice and not play sarcastic pranks. Especially because sometimes I totally miss that comments are sarcastic.

    3. Mario Cantin

      Smart astronaut 🙂

  2. Joe Lazarus

    How does it work? It sounds like it creates a new channel that is tied to a Google Sheets. Does every message in that channel become a row in the Sheet? Can you distinguish between a new list entry and a comment / discussion on the entries?

    1. Joe Lazarus

      Works like this:1) type “/brainstorm brainstorm-name” in any public channel and the bot creates a new channel named “brainstorm-name” and a new Google Sheet with the same name2) type any idea message in that new channel starting with a “-” dash and that idea becomes a row in the Sheet3) type a comment message without a leading dash and it’s not posted to Sheets4) append a “;” semi-colon after any idea message and the bit after the semi-colon becomes a details column in the Sheet.

      1. Alex Godin

        Yup! Let me know what you think. Would love any feedback you have Joe.

  3. Brandon Burns

    Nice bot.That said, we all know that building a business that depends on one specific platform for its success is bad. Everyone here remembers what happened to Zynga after Facebook games stopped being a thing. Yet, nevertheless, the investor camp is calling for bots everywhere! It’s the new future! Forget that building a bot, by definition, makes you dependent on someone else’s platform!Where’s the wisdom in not remembering that history repeats itself?I haven’t commented in a while, and hate to jump back in with something negative, but still. Needs to be said.Again, though, nice bot Alex Godin.

    1. JimHirshfield

      Doesn’t sound like this is a business. But for bots that are, build the bot to operate on all popular meaning platforms, just like apps across many OSes.

      1. Brandon Burns

        Could this bot be built on Kik or Snapchat or another messaging app? And if it could, would people want to use such a bot in those places? Most likely no.Alex maybe built this bot for fun. He also could be reading AVC and other investor blogs, trying to chart his entrepreneurial path, and think that he’s stumbled upon “it.” He’d be wrong.That said, this could very well lead to whatever “it” is, eventually — and that’s why we experiment and build things, and why I’m happy to see this bot built.Still, I’ve just seen a lot of examples recently of the investor community clamoring for things they know are bad. The bot thing gets me particularly lit, because obviously the bots help make their existing investments in Slack, Kik, etc. more valuable, because they’re basically open sourcing their growth — but at the expense of the bot builders who, at the end of the day, mostly won’t benefit from it.

        1. JimHirshfield

          My bot agrees with you more than your bot agrees with you.

          1. Brandon Burns

            Ha! Where’s JLM?!

          2. JimHirshfield

            Eating bots and biscuits for breakfast.

          3. pointsnfigures

            A bot feeds him biscuits and gravy every morning so he can get some south in his mouth.

          4. JimHirshfield

            That’s no way to talk about Mrs JLM

          5. JLM

            .Let’s be careful here because, as you may know, I am an ardent advocate of the reinstatement of dueling and am willing to provide free samples.http://themusingsofthebigre…I thought that Ted Cruz should have called out Donald Trump when Trump insulted Ted’s wife. I would have. Actually, I would have attacked Trump at the next debate and put a thumb in his eye.Sorry, a little overboard. Still, you’ve been warned.JLMwww.themusingsofthebigredca…

          6. JimHirshfield

            Moi? I was protecting the dignity of the missus.

          7. sigmaalgebra

            IIRC, Cruz won Utah because his PACs ran lots of ads showing a GQ cover photo, very professionally done, of model Melania reclining, not very exposed but with no clothes visible. So, somehow, from Trump or someone else there was a Tweet with side by side pictures of the two wives where Melania looked much prettier. So, it appeared that there Trump “counterpunched”. So, the one calling for the dual should have been Trump before Cruz had his reason to “call out” Trump.Apparently at the time, from the delegate counts, Cruz knew he was about to lose and was desperate and willing to try nearly anything to win. Maybe the only reason he didn’t get a gun is that he was too afraid of something or other. He continued that desperation, e.g., with some of his outrageous statements and Carly, and, finally, lost Indiana, IIRC, 57 to zip and “suspended” his campaign.His acts of desperation showed that his view of what is right, ethical, moral, acceptable, appropriate, effective, and good judgment is just what he believes he can get away with legally. Gads. All healthy first graders know much better than that. Looks like the guy was always in the little boys room on the pot when lessons in socialization were passed out. The guy’s social and psychological capital accounts are deeply in the red.From some Cruz statements in the last few days, he’s wound up, all wound up, still wound up, 100% “all in” determined, single minded, charging, never to be deflected, on his apparently “ultimate concern” (P. Tillich) and his likely holy mission for his conception of saving the world with his 100%, unflinching, totally rigid, rock solid, granite hard, iron clad, rock-ribbed, worshiped True Conservative Principles, e.g., a “flat tax”, maybe “The law. The law. The letter of the law.”.Gee, to understand this guy, start with the Diagnostic and Statistical Manual (DSM).But make sure he never gets within 1000 miles of DC and, in particular, is never considered for the SCOTUS.And, just make sure he stays in Texas, hopefully the Texas desert back country. Say, have his PAC build him a nuke-proof, underground bomb shelter to live in.I just feel sorry for his wife and children. They will have to suffer unjustly.

        2. PhilipSugar

          He did sell Dispatch which raised money to Meetup.

    2. PhilipSugar

      I agree when it comes to the consumer. I also agree that being dependent on one platform is really risky.But here you are talking BtoB. There have been many successful businesses built on providing consulting and some technical expertise on Salesforce.It seems (from a quick look) he is building a business to help companies optimize their Slack implementations. Is that a business?? Well yes I could see how if you could prove to companies you could make that more efficient that is viable. There are a ton of people at a ton of companies working on Slack and therefore making them more efficient could have a really good payback. His tagline is “Stop using Slack like an amateur”. I like that.How about this bot then??? Is that a product or in fact is it a REALLY clever way to market his main business slash-hyphen?? I would argue the latter, and getting Fred to Blog about it and getting people to download it is a great way to find the buyers of your other services.

      1. LE

        Really what it boils down to is simply getting a beachhead and established in a way so you can segue into something else. A new product or service (once you have a team) or sell the team and get a good job elsewhere. Sort of what Alex did with Dispatch (as you noted) being obviously acquihired by meetup. [1] See this is what is often hard adjusting to even when you see it happen. The other dimension. Where the net is so vast a failure turns into a success. Meanwhile we see the other day here on AVC that a limited partner who merely passes on a raise by a VC is excluded by that VC in future funds. Maybe everyone doesn’t but the impression I got is it’s fairly ubiquitous.I wonder if they teach any of these “games” (and I say that fondly with admiration not as a pejorative) in business school.[1] http://techcrunch.com/2013/

        1. PhilipSugar

          I think what they don’t teach is sales. Now many people think that sales is just “conning” somebody into buying your product. That is not the case.And many people think that marketing is just talking about your product. To be good at either you can’t have that mindset.I love the fact that in reality what he did is to add value when he was marketing. That is the best type. That is in fact what Fred is doing here on this blog.I just bought a part for my boat…..look at this guy’s writeup explaining what the problem was and why he had a better solution. http://www.sbmar.com/produc

          1. LE

            Agree with p1 to p3 but disagree with the link as far as marketing.First agree that these are “goodfellas” http://www.sbmar.com/the-se… seems like a solid crew that knows and loves this shit. So a win there.But the info for that product isn’t organized and rambles on to much. Maybe your brain likes that (it does that’s your point) but for my brain it doesn’t pass the puny brain test. Maybe if they are selling a product to an engineer at boat manufacturer who has time to read something so involved but the product is a $300 product seems way overkill. The overkill is fine actually but needs to be presented in an easier to digest manner and with more clear headers so people can focus on what is important to them. So the “crime” here is bad presentation. To much good stuff requiring reading. Overselling. Presenting tons of info is great but it needs to be done in a way that doesn’t make you think “I am out of time”.By the way talking in detail like they are doing is a form of “conning” in the sense that the people you are selling to often don’t know enough to know if what you are saying really matters or not. It’s a technique that is often used also by politicians that are policy wonks and leave normal people in the dust.Ok, here you go, all the excess info (other than maybe 1 paragraph summary) belongs on a page like this linked (on both pages):http://www.sbmar.com/articl

          2. LE

            By the way the crew at Seaboard just goes to show there are plenty of things people can do to earn a living w/o having to get a 4 year liberal arts college degree and then work waiting tables or doing something they dread. The guy at marine max that I used to deal with started as a mechanic, went on to run the service department, and now is running an entire region for the corporation when I checked a few years ago out of curiosity. My ex brother in law dropped out of college to join his father’s alarm company (which I at the time thought was a big mistake for other reasons). He is now running the company and (starting doing installs) and has a pretty nice life. (The father never went to college either he started as a tin man in sales and went from there.)

        2. sigmaalgebra

          I wonder if they teach any of these “games” (and I say that fondly with admiration not as a pejorative) in business school.My wife was ill, and to have a way to help her recover, e.g., have more time for her and to have her near her family farm in NE Indiana, I took a job as a B-school prof at Ohio State in Columbus. While my wife didn’t get better, I did learn some of how B-schools worked.I expected that B-school would be professional like law or medicine, e.g., have a clinical side where business people with real problems would come for real solutions and students would get to see the work, be apprentices, etc.Nope: Instead the main theme that was admitted was that B-school was to be applied social science like engineering was applied physical science. The real theme was that they had a bad case of physics envy where they wanted a business version of Newton’s second law F = ma or Einstein’s result of his special relativity E = mc^2. E.g., they wanted to look at investing as just modern portfolio theory building on Markowitz and Sharpe. For building on Markov processes and potential theory, they were missing about six years of math major focused course work. E.g., they were much more interested in far-out mathematical efforts to chip away at the question of P versus NP instead of showing people in business how actually to make valuable progress on problems in production scheduling, logistics, vehicle loading and scheduling, and capital budgeting. Indeed, any contact with real business was regarded as unprofessional and unethical.I never wanted to be a prof at all, and when my wife wasn’t getting better, I said we needed to have us go to where I’d have a chance of getting a real career going.So, for teaching “games”, nope!Of course, some of the students were from families successful in business and had, since middle school, been learning about the family business. So, for a course in, say, labor law such a student, having participated in labor negotiations for his family business, could easily know a lot more about labor law than the labor law prof.E.g., typically a prof of business information systems was missing any significant hands-on computer knowledge, knew next to nothing about programming languages, SQL, management of software development teams, etc.E.g., when I was in grad school, after several years since college with a nicely successful career in computing (annual salary about six times what a new high end Camaro cost), there was a course in computing I was urged to take. So, I did. I’d already taught several sections of a comparable course at Georgetown University.Well, in that course one of the profs, actually the best of three trying to teach the course (I’d taught my course alone as a side project while being the main technical guy in the computer center), had his career based on some tricky progress in some niche aspects of string search. Okay. But near the end of the course, he was to spend the summer helping some people doing real computing and asked me if I knew JCL. Sure I did. That was IBM’s mainframe Job Control Language. Actually the language was fairly simple if well explained, which it nearly never was, with just frustratingly delicate syntax and infuriatingly obscure semantics, a total unanesthetized root canal procedure with “a barbed wire enema”. Actually much of the language was just a way to fill in a control block data structure for some assembler language macros for doing input and output. So, really, the main documentation was just that for the assembler macros, but this point was rarely made clear! Could spend the whole summer just mud wrestling with the syntax and the cubic foot of obscure documentation. Have a nice summer, guy!Lesson: Nearly all of research university education in applied math, applied science, engineering, and business doesn’t want to be practical or even professional and often in significant ways falls behind the more advanced parts of practice. Really, commonly the researchers are just too far from the real world to pick good research problems. Sure, Einstein could sit in a small room and dream up E = mc^2, which is also fully terrific for a lot in practice, but stands not to happen very often and is not nearly all there needs to be to research, teaching, and practice in applied math, applied science, engineering, and business.For my startup, I took advantage of the best I did learn in grad school at a world class research university. Well, in some major ways, that is a good thing to do: The pure math prerequisites for the crucial core, original applied math for my startup was good stuff decades ago and will remain good stuff decades from now! Meanwhile, the more practical stuff then is obsolete now! So, for the more practical stuff now, that’s all conceptually just trivial; I have to teach it to myself; and the main bottleneck is that the computer industry gets a grade of D- in technical writing.Law profs continue to practice law. Med school profs continue to practice medicine. But schools of applied science, engineering, and business don’t want such practice.

      2. Alex Godin

        Exactly Philip! You just nailed the business plan…YesAnd is one of a few bots we’ll be releasing this year to help lay the groundwork for our consulting business. When we saw Fred’s post, we knew we had to build his bot.

        1. PhilipSugar

          As you can see I think it was a brilliant move.

        2. Lawrence Brass

          Doing consulting and product development sounds perfect, best bootstrapping strategy around. A smart consultant can get a lot of input viewing and analyzing customer processes and flows that can be represented later in a product or a ‘core solution’.Great work.

          1. Alex Godin

            Thanks Lawrence!

        3. Anne Libby

          Go, Alex.

      3. Brandon Burns

        Agreed (mostly).Alex built a cool bot, and it’s already gotten his main business some press. These are good things. Bravo, Alex.Still, nary a day goes by that I hear some disillusioned budding entrepreneur tell me some idea they want to build in response to the investor community’s latest whim. Now we have all these people trying to build bots, or AI stuff, or VR stuff — and the “solutions” are mostly terrible. “But that’s what the investors like!” they say. Sigh.So when I read that this bot was built in response to a request from Fred (which I don’t think is bad, in and of itself) I think about the larger message that it sends the many impressionable readers of this blog who are trying to find their path towards startup gold.Again, great job Alex. That said, the success of this experiment is more exception than rule.

      4. Mark Essel

        Well said.

    3. Alex Godin

      Thanks Brandon! There are definitely risks inherent in building on top of a product like Slack or Facebook. At the same time, there’s a huge amount of power that comes with the Slack platform. Apps that build on top of Slack can do things that would be almost impossible for a young upstart.In our case, we’re focused on building a consulting firm rather than a product company. Our model is closer to Bluewolf and Silverline (salesforce consultants) than Zynga or Branchout.

      1. Brandon Burns

        Kudos on a nice product.Everything else I have to say, has already been said in response to Philip. 🙂

        1. PhilipSugar

          I hope you understand the reason I upvoted your comment was that it made me think, and I believe that a civil banter is greatness, and I know you do as well. Best regards, my friend. 🙂

          1. Brandon Burns

            “civil banter is greatness”I know some adult-aged children who could hear that statement :-)(not talking about anyone on AVC)

    4. falicon

      I wrote a post a few weeks back with this same basic stance, bots are the new API -> http://falicon.com/post/142

      1. Brandon Burns

        Nice! On today’s reading list!

  4. mikenolan99

    Just posted to our Slack channel… 🙂

    1. fredwilson

      Hehe. Nice

    2. Alex Godin

      Nice! Let me know what you think Mike.

  5. markslater

    We are trying out the Slack <<>> Asana hook up.

  6. jason wright

    a bot complements an app like Sheets, and then replaces the app?a decentralised bot wins?i know little about bots yet.

  7. Matt Zagaja

    So far this is the only bot that I want in my channel:

    1. falicon

      Speaking of which, when is the FAKEGRIMLOCK bot going to be released? ME WANT TOO.

    2. ShanaC

      I miss him

  8. Jess Bachman

    Side note, I recently switched from Slack to Ryver. It’s just better, in just about every way. I give Slack credit for modernizing IRC, but the innovation beyond that has been lacking.. oh right… super large emojis. Ryver.com has solved some major paint points for me.

    1. JimHirshfield

      Which exact paint points has it solved for you? Cleaning your paint brushes?

      1. Jess Bachman

        The concept of threaded conversations. With slack, I post a link about something.. some people discuss it right then… then the convo moves on in the channel… then you can’t discuss it again without feeling like you are bringing up something old. Ryver has a forum like mechanic, in addition to straight chat, which works perfectly for this.Oh.. and its also free.

        1. JimHirshfield

          Good point. That is definitely something I experience on Slack.

        2. Alex Godin

          Ryver sounds really cool. I know that Slack has been experimenting with threaded conversations for a while internally. Hopefully we’ll see something from them on that front soon.Do you know about Ryver’s funding status? I’m somewhat wary of a 100% free workplace app.

          1. Jess Bachman

            $9.2M. Good enough for me. http://www.marketwired.com/…I agree with their pricing premise tho. Slack is just prohibitively expensive, especially for smaller unfunded teams. It just adds up.

  9. Rob Underwood

    I had my money that the next bot you’d feature would be the port of the game Zork I did to Kik (bot name is ‘zorkgame’) over the weekend. 😉

    1. fredwilson

      i have to play it first

      1. Rob Underwood

        cool. just rebooted the node server. up now.

  10. pointsnfigures

    The “Yes, and” bot. Maybe it can disrupt Improv.

    1. Alex Godin

      Ha! We leaned on improv for naming inspiration.Credit where credit is due, @ruthienachmany helped come up with the name.

      1. ruthienachmany

        🙂

      2. Kirsten Lambertsen

        I was wondering if that’s where the name came from! Cool 🙂

  11. Lawrence Brass

    I am looking at bots now as a form of user defined loosely coupled system integration. It brings up some memories of power users and Excel macros doing magical things. A few times I got contracts where the definition of the problem and the (unstructured) solution were contained in a highly automated spreadsheet. The task was to convert these into ‘compliant’ applications.If this bot thing wants to fly it would be very helpful to have a common bot API including bot authentication. Maybe someone is working on that already.

    1. sigmaalgebra

      General rule of software productivity via extensibility: If you have some software, then design it so that it can be driven by other software.

  12. Robert Rogge, CEO Zingword

    How many people do you need on your team until Slack becomes really useful? We tried it, but quickly returned to Skype. Old habits…

    1. Alex Godin

      I think it depends on your team and what you were using before. We’ve seen some really successful 4 person teams, but Slack really starts to shine at about 10 people.

      1. Robert Rogge, CEO Zingword

        That’s what I figured… the channels would be useful at about that time. With a smaller team, it feels like “what are we doing in here?”

    2. Kirsten Lambertsen

      I belong to a small, informal group that uses it quite effectively. I’d say the issue isn’t the size of your group. It’s more the need for both async and real-time, the ability to quickly share documents, links, with more than one person, and the ability to integrate with your other tools (as evidenced by Fred’s post today).At work, we use it to deploy code. We use it to alert on various things.Also, searchability. Just last week I found an answer I needed by searching Slack history, rather than pulling someone off of what they were doing.

      1. Robert Rogge, CEO Zingword

        Not disagreeing Kirsten. But you’ve got every possible buzzword packed into your comment. Email, Skype, and a bunch of other communication tools are essentially async and real-time, and you can quickly share documents and links with more than one person.Just last week, I found an answer I needed by searching my email history, rather than pulling someone off of what they were doing. It was an incredible experience. :PAnd alerts… alerts are cool. But Trello pings my phone anyway, you know? Do I really need an alerts channel?I think integrations for alerts etc. and clean UI concepts of teams and channels are the main reasons to use Slack. I have fifty-thousand Skype contacts, and no matter what I do, it never feels like there is a place just for us. We’ll probably Slack again when our team is bigger because it’s a great team chat thingy.

        1. Kirsten Lambertsen

          Sure, I mean, I’m not here to sell you on Slack. But I still don’t think size matters. I’ve used it as part of a large-ish company and a small informal group.The difference between searching email and Slack history, though, is that Slack is more like you have access to the entire team’s email archives, not just your own. (The answer I found was a conversation that took place between two other people *before* I joined the company.)For me the biggest benefit to Slack is to get communication and knowledge unlocked from personal email silos (oh sh*t, a buzzword) etc and into a place where it can be of more use to everyone involved.

          1. Robert Rogge, CEO Zingword

            Not trying to hate on Slack either.Salient points, all of them. For a small team though, you need to *want* to change from whatever chats you are already using. That searchable history isn’t really useful until later when you need it. And most of the conversations are us talking to ourselves, anyway.I could do without that word “silos,” Kirsten. Because I’m pretty sure we were talking about email “accounts.” :POn a fundamental level though, Slack changes the concept of “mine” and “ours,” where your conversation is actually “ours” to search, like in your case. I think if you were at an office gig in the ’70s, people would freak out if you told them that from now on, the rest of the office could read a transcript of their meetings. 😀

          2. Kirsten Lambertsen

            Sure, again, I’m not invested in you using Slack ;-)But I’m going to defend the word “silos.” It’s used to create a more vivid image of the situation — email being silo’d in personal accounts. Granted, its current status might be as a buzzword, but it still does the job I hired it to do.Zingword looks cool. Hope it’s going well!

          3. Robert Rogge, CEO Zingword

            We’re doing great, thanks for the kudos. We have a 3 person founding team including a product designer/ux, which is why our site is so sexy.We’ll launch ahead of a competitor who was recently acquired, and we’re doing it better than they are because we have e-commerce, design, and sweet developing chops in the silo.Then, we’ll need a round of capital to finance the race. It wouldn’t surprise me if we don’t get offers shortly after launching because of some of the other acquisitions that went out lately – we will be in the lead, and anybody who wants in will need to do that in 2017.That’s how I see it anyway. I don’t get out much these days. Feels good to talk on Fred’s comments section. :PYou can DM me on Twitter.

  13. falicon

    So much AWESOME! I love it when the community builds something based on your posts too! Yes and, it’s official, I’m the newest super-fan of Alex! 😉

    1. Alex Godin

      🙂 Thanks!

  14. sigmaalgebra

    I’m lost: So, the need is a way to handle ideas and lists of ideas? And spreadsheets are regarded as a good way to do that? And for a group of people working collaboratively on the list of ideas, want a bot communications interface to the spreadsheets?In simple terms, I thought that on-line fora (forums) were good for working collaboratively on ideas, say, from Disqus and maybe something from WordPress back to fora in the days of dial-up to AOL, CompuServe, Yahoo Groups, etc.And another way is just to have people get remote access with a directory of an hierarchical file system with, sure, capabilities and access control lists.Really, just the default basic hierarchical naming that goes with HTTP and standard Web server software functionality that makes the HTTP hierarchical names correspond one to one with the file names in an hierarchical file system should be a good start. For some added functionality, just put some Web site code between the HTTP and the file system, that is, a case of essentially what all Web sites do anyway.Beyond the communications plumbing, I’m lost on why anyone would find spreadsheets a good way of storing and presenting ideas or lists of ideas.Instead, I always thought that the main way to present ideas was in the usual means of preparing, say, documents from short notes of a few lines, to longer notes in e-mail, to posts such as in Disqus, to whole Web pages in HTML, to Web sites with lots of Web pages in HTML with links for cross references, to such Web sites (e.g., Microsoft’s MSDN) with a hierarchy of such Web pages, to foils, to papers, maybe in PDF but possibly HTML, with title page, table of contents, chapters, sections, subsections, cross references, glossary, and index, to collections of papers, to books, to collections of books, etc.For the role of spreadsheets? I’ve used spreadsheets, but, really, for presenting and organizing ideas and lists of ideas? What am I missing?For lists of ideas, about essentially everything of importance to me, I have a file that I use and maintain with just my favorite ordinary text editor, in my case, KEdit. The file has the idea itself or pointers to other data with the idea.The file is currently just over 2 million bytes, has 66,000 lines of just simple typing, was started 11 years ago, and has 3600 entries. Each such entry starts with a reserved token :Keys. and, then, has, let’s see, on average, 14 lines. Each each entry starts with a delimiter, a line of characters ‘=’, a line with the reserved token :Keys, with keywords, a blank line, the lines of text, and a final blank line. Simple. The keywords should have synonyms, but I have not seen a big need for that feature and have not implemented it.So, the file has phone numbers, mailing addresses, notes on people and organizations, links using file system tree names or Internet URLs, etc. There are ideas possibly useful for my startup, favorite music, e.g., with YouTube URLs, good quotes, some from books, movies, famous people, and more.For the goal of real artificial intelligence, (Ai) it is astoundingly, overwhelmingly shocking just how small the file is considering how important its contents is for me. So, candidateAI Lesson: A key to real AI is good knowledge representation. One reason is the promise that, with representation as efficient as my little file, a trillion bytes of data might be enough for one heck of a powerful case of real AI. A sufficient key to such representation is, clearly, just good natural language understanding with just text input and output.For the understanding, that is likely essentially just an all internal, self-referencing situation in the sense that the main criterion is just that the AI can work according to its own internal goals with its own internal techniques with just its own internal data and, of course, also do meaningful natural language communications via text input and output. That is, in a significant sense, the necessary meaning and understanding of real AI are, in some architectural sense, a bootstrap situation, that is, mostly just a case of self-referencing or, if you will, just an internal consistency, instead of something more obscure and complicated.

    1. Stephen Voris

      Perhaps think of spreadsheets of ideas as analogous to poetry?Where the structure of the words lends meaning as well as the syllables.Not quite so structured as code, nor in this case associated with art, but still useful. A form(at) of metadata.

  15. Donna Brewington White

    Post and you shall receive.

  16. ShanaC

    Also this is a frustrating post, because the one “bot” or “ai” I want to talk about I’m not sure I’m allowed to.Boo