AI on a blog

by Leave a Comment

Reading Time: 3 minutes

One of the areas of AI that I’ve very interested in is using it to comb through my own information. I’ve used it a bit with my Obsidian notes, and more tools like that are coming out all the time.

While I’m using that for personal use, what if I could give you access to an AI-built version of my brain? A good place to start could be on this blog, which has around 1,500 posts going back almost 20 years. It seems like it would be neat, but there is a big downside.

Seth’s AI

Seth Godin is playing with this idea, and you can go search his blog using AI right now. It generates a list of relevant posts, and then writes a summary of what his thoughts on the subject likely are. It’s nice that it keeps the list of posts visible so you can dig into Seth’s actual words to see what he said, as the summary from “him” may or may not accurately reflect his thoughts.

Change your mind?

Where this could get troublesome are for topics where I’ve changed my mind over the years. I post frequently about changing my mind as I gain more insight, but would the AI know that? It seems that if you asked a question about one of those topics, AI might report “my” response as being my old thoughts on that idea, and not representative of my current state at all.

That’s kind of the philosophy that Fred Wilson has taken with his blog. He has many thousands of posts on his blog (far more than I do), and he’s specifically not offering an AI solution to “ask Fred”. If you have a question for him, he’d prefer that he gives the answer himself, rather than AI taking a shot at it.

Anyone can

The problem for Fred (and for me too, I suppose) is that we don’t have the final say. While neither of us have an “ask with AI” option on our site, all of our content is public and indexed and someone else could write a chatbot using our content if they want. I don’t know the legality of that, but the technology to do it would be pretty straightforward.

This will lead to some very interesting things down the road where you can essentially ask a person a question and get AI to summarize their likely response. For someone like Seth, who has thousands of posts, videos, podcasts, and around 20 books published, there is a ton of data to feed into the system.

We’ve already seen a bad example of that where a magazine published an “interview” with Michael Schumacher, but Schumacher’s quotes were all made up by AI. It’s a very interesting idea, but it was handled incredibly poorly and the editor of that magazine has since been fired.

Chat with anyone

Seth also published something very interesting on his podcast as well, where an entire episode was written and spoken by AI. I’ve listened to most of his shows, and I didn’t catch the fake — it sounded just like him.

With that, it’d be trivial to build a system that let me call in and have a phone call with “Seth”. It wouldn’t be him, but it’d sound like him and give answers that are likely similar to what he’d have to say.

As with most things AI-related, you can keep walking this trail forever. Why not have a video call with this fake “Seth”? Or have him call in and be on an interview panel with a mix of real and AI-generated guests?

There’s some fascinating stuff coming, and your best bet is to not ignore it so you can understand what your options are in the future. If you’ve not substituted many of your Google Searches for ChatGPT searches, simply to play with things and test out ideas, that’d be a solid place to start.

The lock icon in your browser is going away because people don’t understand what it means

by Leave a Comment

Reading Time: 2 minutes

Web browsers have had the padlock icon next to the website addresses for years to show that the site has an SSL certificate and that your data is protected between you and that website. For years it was mostly just ecommerce that had that, but in 2017 Google started pushing all sites to use SSL.

SSL is a great thing, but it’s often very misunderstood. At a basic level, it protects that information that you give to a website (so no one can see it in transit), but that’s about it. It offers no other protection for the website itself, or for your assurance that the website is legit. As I heard a friend say at a security conference, “SSL just means that hackers have a secure way to get your site”.

As a consequence of the confusion, many users see the padlock and assume that the site is completely safe — not only is their data safe, but the site itself is reputable. Those two things are completely unrelated, and it’s led to problems. Here is what Google had to say about it:

Despite our best efforts, our research in 2021 showed that only 11% of study participants correctly understood the precise meaning of the lock icon.

This misunderstanding is not harmless — nearly all phishing sites use HTTPS, and therefore also display the lock icon.

Any good scammer will use SSL, so that icon is of no value. Google has slowly been making the lock less obvious, and now they’re going to replace it with this new symbol:

While I think it’s an improvement in some ways, as people won’t see the misunderstood lock, it’s a confusing icon itself. Personally, I think Google should have just done away with that icon completely, and only alerted users when a site was not using SSL at all.

Either way, this should help a little bit, and it’s a reminder to all of us that hackers do many things to try to appear legit so always be on guard.

For more, you can read Google’s official post on the change, or this great summary from Search Engine Journal.

Don’t freak out about AI just yet

by Leave a Comment

Reading Time: 3 minutes

I’ve shared a few times on here some concerns I have with AI. There are good things coming, but probably some bad things too. You’d be wise to spend time with AI tools and start learning how they work, but it’s not time to freak out over it yet.

Napster

Back in 1999, the record labels were freaking out. Napster was the cause of it, and perhaps rightfully so, but the labels let their fear spread into weird areas.

With Napster, the record labels essentially did nothing other than try to stop it. No innovation, no new ideas, just lawsuits. Fortunately for them, Apple and Pandora (and later Spotify and others) found ways to make them money in the new world.

OLGA

Around the time that this was going, there was a site growing in popularity known as OLGA: The OnLine Guitar Archive. It was huge database of guitar chords so you could learn to play your favorite songs. While their use of lyrics on the site was indeed illegal, it was a weird fight for the record labels to make, since it wasn’t competing with a paid alternative. The labels chose to fight anyhow, and they got OLGA shut down.

Around the same time, I ran a very popular Christian music guitar site with similar content (lyrics + chords to popular songs) and faced the same issues from record labels. I sold the site years ago, and it finally shut down last year (2022).

From this first-hand view, three interesting things happened while running that site during the Napster days:

  1. In speaking with a record exec on the phone once, he admitted that our site was good for their business. Churches pay license fees for contemporary music on Sunday mornings, and we were feeding into that and helping their bottom line. Still, he said he “just couldn’t let us keep showing the lyrics”, but with no reason why.
  2. I also would sometimes showcase new music on the site that I enjoyed. I wouldn’t share the music itself, but just talk about an album and show a photo of the album cover with a link for people to go buy the CD. With around 25,000 daily users, it likely sold a good number of albums for them. I was told to remove the pictures of the album covers because they were under copyright. 100% legal for them to do that, but mind-numbingly foolish.
  3. In terms of lyric issues, I simply removed the lyrics from the songs on the site, which made them 100% legal but 99% less useful. Instead of seeing which chords to play as the song went on, you’d just see “This song is mostly D – G -Em – C.

Because the record labels were scared of MP3s and Napster, they just went around trying to stop everything, even if it was to their detriment.

Cameras

Around that time, another interesting thing happened — concert venues disallowed cameras. I would often take photos at shows and share them on the site (thus sharing this artist with thousands of new users), but they decided that was a bad thing. Again, no good reason why, but just a decision that it must be stopped.

Of course, that ban ended a few years later when literally everyone was walking around with a camera in their pocket.

AI is feeling the same

While AI is a very different situation in a lot of ways, reactions are quite similar.

People are scared of what it might mean.

People are scared for their jobs.

Schools are banning it.

Like the rise of MP3s, AI can’t be stopped. It’s coming, so get educated and learn how to use it for good. Record labels freaked out over MP3s but did literally nothing to try to harness the power of them. They got lucky that Apple bailed them out with iTunes, but your industry might not be so lucky.

Are you worried that your job might be taken by AI? Do better.

Not “do better at your current job”, because you can’t outrun AI forever. Instead, “do better at what your job will become” by harnessing AI to multiply your skills and output in a way that can’t be replaced.

If you’re not sure how to “do better”, I recently heard a great tip — replace Google with ChatGPT for a week. All of the things you might ask Google, ask ChatGPT instead. Some answers will be better, some will be worse, but you’ll learn a lot along the way and have a much better idea of how these new tools can benefit you.

Using ninety.io with Traction

by Leave a Comment

Reading Time: 2 minutes

I’ve mentioned a few times in the past how we use the EOS / Traction model at GreenMellen. If you’re not familiar with it, it’s a way to organize and manage your business (with “EOS” standing for “Entrepreneural Operating System”) and the book “Traction” is a great place to start.

Funny enough, my main post about EOS is how we don’t follow it too precisely. Our business is now at a point where we need to focus on a bit more precision with it so we’re digging in further.

To this point we’ve managed most of it through Notion with our team, including our meeting cadences and issue lists, and Google Drive for our scorecard. My plan was to continue using those systems, but I didn’t realize that there are digital tools out there that are specifically designed to help manage your Traction processes, and a great looking tool is Ninety.io.

Ninety

We’re just getting started with it and our thoughts might change, but so far it seems like an amazing tool. They include pieces for running your meetings, tracking performance and goals, some light task tracking, an accountability chart and more. It’s all the same stuff we’ve done elsewhere, but in a single place with ways to connect all of them (like turning an “issue” into a “task” with a single click).

It’s only $16/user/month, which seems pretty fair to me. We’ll only have a handful of people using it, so it should stay under $100/mo total. If it can really help us tighten up our use of Traction, it will pay for itself 100x.

If you’re an EOS/Traction user, I’d love to hear if there are other tools that you prefer.

My 486 wasn’t powerful enough, and neither is AI

by Leave a Comment

Reading Time: 3 minutes

Back in the mid-90’s, I worked for a few years at Electronics Boutique, a software and video game store. It was great! It was a fascinating time in the world of video games, going from systems like the Super Nintendo and Sega Genesis, through failures like the 3DO and Virtual Boy, into the next generation with the Nintendo 64 and the first PlayStation. It was such a fun era to be working in a store like that.

Computers were also moving quickly, and I still remember a conversation I had with a woman that came in to shop. She had just purchased a 486 DX2 66, a very popular chipset for a few years. At the time, it was top of the line and could run any program or game we had in the store. In fact, it was seemingly so powerful at the time that she said “the power of this computer will last for the rest of my life”.

Having seen the progress of computers to that point, I knew she was completely wrong and that by the next year we’d already have games that she probably couldn’t run. “The power of this computer” wasn’t likely to last more than a few years, much less the rest of her life. It’s easy to get excited about the powerful new thing we have, but history shows that technology becomes obsolete remarkably quickly.

Case in point, the latest iPhone has roughly 650,000 times the computing power of that computer.

5 gig fiber?

This same thing has progressed as time goes on with storage space and speed, and we always need more. Right now I’m seeing it with internet speed, and “why would anyone need 5 gig fiber?”. In the short-run, they’re correct. I have 1 gig fiber and it’s fantastic; I see no reason why I would need 5 gig, but I also have no doubt that I will in the coming years. Perhaps it’s due to VR, or more 8K content, or a combination of things, but I’m completely sure that we’ll need increasingly fast internet speeds in the future. I can’t say specifically why we’ll need 5 gig fiber, but I’m entirely confident that we will.

Now it’s AI

AI will follow the same trajectory.

  • I hear plenty of folks saying that “it’s easy to spot content written by ChatGPT”, and they’re right.
  • The AI-generated images from Midjourney are imperfect, especially with things like hands.
  • The AI-powered Bing search results are amazing at times, and stunningly awful at others.

This will all improve very quickly.

I can’t tell you what these engineers will do specifically, but this stuff is improving at a staggering rate. I can promise you this: your next computer will be faster than the one you’re on now, your next home internet package will be faster than you have now, and the amount of data behind the AI tool you use will be massively larger with each upgrade.

Multiply that by 10 or 20 years, and it’ll be a revolutionary difference.

If you think you have enough hard drive space for the rest of your life, or if that shiny new fiber connection is the fastest internet you’ll ever need, I think you’re wrong. I can’t say what exactly will cause those to become insufficient, but history has shown for decades that it will, and I’m excited to see what new technologies come along to make those upgrades necessary.

As automation speeds up, relationships will reign

by Leave a Comment

Reading Time: 2 minutes

I recently shared a glimpse into the “bad side of AI“, where I’m starting to get some automated comments on social media and it’s likely to get much worse.

The side I fear a bit more is the coming glut of AI-personalized email spam, at huge scale. Microsoft has unveiled Viva Sales and Salesforce announced Einstein GPT, both of which use generative AI to help further automate the sales process via email.

Salesforce isn’t even pretending it’s something else, as one of the features they tout is that “Einstein GPT can generate personalized emails for salespeople to send to customers“. Horrible.

As I’ve shared before, cold outreach is the domain of spammers. It’ll be a little trickier to detect, as these AI-generated emails will probably become quite convincing, but it’s unsolicited spam no matter how you try to spin it. Building relationships, which have always been the key to solid partnerships, will reign.

As Gary Vaynerchuk recently said, “the advantage becomes the things that are not automated… that are human-based, and this will be accelerated dramatically with AI“.

AI can still help

I’m still excited for how AI can help in these conversations and I see a few great ways to incorporate it:

  • Automated note-taking. Google has tools coming soon that will help take notes during video calls; taking that off my plate will allow me to be even more invested in the conversation and not have to worry about capturing everything myself.
  • Prior to meetings, AI can help summarize those notes can get me up to speed. We’ve already seen a tiny peek at that, and much bigger things are coming in the next few years.

That said, the amount of semi-personal spam that I get today is already wildly unacceptable, and it’s going to get worse. However, as it gets worse for everyone, the effectiveness will continue to drop. The replacement will be tighter relationships with those around you, which has always been the ideal way to do business.

Powerful systems that work come from simple systems that worked

by Leave a Comment

Reading Time: < 1 minute

Nearly 14 years in, our process for building websites and digital marketing strategies is pretty fantastic. The outcomes, both visually and financially, are a winner for everyone involved.

14 years ago the process wasn’t nearly as strong or detailed, but it worked, and helped lead us to where we are today. In his book “The Personal MBA“, author Josh Kaufman shared a short statement that got me thinking about this, know as “Gall’s Law”. It simply says:

“all complex systems that work evolved from simpler systems that worked”

If you went back in time 14 years and handed Ali and I our 2023 process for marketing, it wouldn’t work. We didn’t have the talent or the staff, and to a large degree the software in 2009 wouldn’t have been sufficient. Even if you hand our process to another agency today (which we often do), it won’t work for them because it’s been fine-tuned to our ideas, beliefs, and talent. That’s not to say our process is better than what another company can do, but just different.

Early on, Ali and I worked hard on building processes to help ensure success, even though the processes were quite simple at the time. From that simple process though, we’ve been able to evolve it to a complex series of steps that works wonders. Without that initial base, we wouldn’t be what we are today.

Most of you likely have written processes in place for what you do, but if you don’t, now is the time to start. Get that simple system humming, and as you slowly refine it over the years it will turn into a thing of beauty.

Generative AI is better for input than for output

by Leave a Comment

Reading Time: < 1 minute

AI is all the rage these days, particularly “generative AI”, or AI that can create new content. This could be new text output, new images, or even new videos. The power is amazing, but I see two very different sides to it.

Input

I love the thought of how IA can help with input. I can have it summarize podcast episodes, dig through my notes, or boil a complex topic down into terms that I can understand. It can help a lot in education, and there are huge implications everywhere.

Google has some new tools coming soon that can take notes for you on video calls and summarize long email threads, and I think both of those sound very compelling and likely quite beneficial to all.

Output

The other side is output, and people using AI to generate new content for others. There are certainly cases where this could be beneficial, but most cases lean more toward spam.

You have things like the new IFTTT tools I mentioned a few days ago, and part of Google’s new update is to allow AI to craft emails for you.

Of course, you can’t have one without the other; if AI is powerful enough to summarize data for me, it can certainly do the same with other content and just tweet “new” content all day long.

I suspect both the input and output scenarios will continue to grow in the coming years, but I’m holding out hope that the benefits of AI-powered input will heavily outweigh the faux social media and blogging that’s likely to come from the output of a generative AI system.

It only takes five seconds to clone your voice

by Leave a Comment

Reading Time: 2 minutes

I remember watching a video last year where some software was able to listen to a recording of a voice for a few minutes and then mimic it with amazing accuracy. Since then, the tools have become more and more powerful and a typical voice model can now be trained on just five seconds of listening to your voice.

Here is an example showing some examples of AI Clones working on just five seconds of input:

The repercussions of this are massive. Matthew Wright shares a hypothetical example:

You have just returned home after a long day at work and are about to sit down for dinner when suddenly your phone starts buzzing. On the other end is a loved one, perhaps a parent, a child or a childhood friend, begging you to send them money immediately.

You ask them questions, attempting to understand. There is something off about their answers, which are either vague or out of character, and sometimes there is a peculiar delay, almost as though they were thinking a little too slowly. Yet, you are certain that it is definitely your loved one speaking: That is their voice you hear, and the caller ID is showing their number. Chalking up the strangeness to their panic, you dutifully send the money to the bank account they provide you.

The next day, you call them back to make sure everything is all right. Your loved one has no idea what you are talking about. That is because they never called you — you have been tricked by technology: a voice deepfake.

To be honest, I’m not sure how to prevent this. Additional security on cell phone number spoofing would be a good step, and it’s coming eventually, but right now this is a very real possibility. If my daughter called, from her number, and with her voice, asking for help right away, I’d do it.

A recent episode of Seth Godin’s “Akimbo” podcast dug into this a bit more, where he used ChatGPT to write much of the script, and then he used an AI voice model of himself to read much of it. It did a stunningly good job with both of those.

The personal security implications of this, like I shared above, are a huge concern but it goes beyond that. Right now, I could ask my daughter to hop on a video call to prove that it’s her and that solves it — for now. In a very short amount of time, that will be easy enough to fake as well.

What’s the solution? I don’t have it, but the people that can best solve these kinds of crazy new problems are likely to be the next set of billionaires that we’ll be talking about in 2030.

Thank you for your time

by Leave a Comment

Reading Time: 2 minutes

If someone invests time with you, whether it’s to help with a project, move a sofa, or just go to the movies, it’s appropriate to thank them for their time. They willingly gave of themselves, and deserve a bit of thanks for it.

What if you steal their time, though? Can you still thank them for something they didn’t mean to give you?

Stealing apples

Suppose you’re at a farmer’s market and see some amazing apples. You could, if you chose to, walk up and take one of the apples and shout a “thanks for the apple” back to the seller as you walk away. It sounds like something a bully would do in cheesy movies from the 90’s.

I’m finding this in some cold emails that I get, too, and I interpret it the same way. Simply the act of opening the email is an unearned theft of my time, so a “thank you for your time” doesn’t make it any better.

Here’s an example of one I recently received:

They have the “unsubscribe” info at the footer, but you should never find yourself in a situation needing to unsubscribe from something that you didn’t subscribe to in the first place. Seth Godin summed that up perfectly in a post from almost a decade ago:

Some spammers will tell you that all you need to do is opt out. But of course, the very problem with spam is that it requires action on the part of the recipient, action that can’t possibly scale (how many times a day should we have to opt out, communicating with businesses we never asked to hear from in the first place?) People are smart enough to see that once spam becomes professionally and socially acceptable, all open systems fall apart.

This is similar to spammers that want to become a “trusted partner” after starting the relationship with spam, and it’s just such a pitiful way to try to begin a relationship.

You should absolutely include unsubscribe links in bulk emails that you send, you should thank people for their time, and you should work to be trusted partners with those you serve. These are all good things, but only in the right context, and none of them make a cold email any less annoying and spammy than it already is.