Florida-based Avalon Corp. has raised $13 million to build a platform for games capable of taking advantage of the metaverse, featuring a universe of interoperable worlds with various intellectual properties, to be built using technologies including blockchain and game engines.
According to the press release, Avalon Corp is uniquely experienced to solve problems that will face creators and designers in the near future, and is building the tools, framework, and more that will allow them to do so. CEO Sean Pinnock has said that the company is working towards the democratization of game creation using blockchain as a tool.
According to Pinnock, it’s gaming companies that should be building the metaverse, not anyone else.
“We know that engagement is driven through play, creativity, and human connectivity, all critical elements that are severely lacking in most self-proclaimed metaverses, and it’s exactly why the inevitable confluence of tech that will emerge as the metaverse will be built by game developers,” he said in the announcement.
He does have a point. After all, the global video game industry is now bigger than the movie, music and book industries combined. And video games, especially first-person, multi-player video games, have exactly the kind of technology — and engagement — that the metaverse needs. And that companies like Second Life, Facebook and Microsoft have, so far, been unable to tap into.
It’s also a good sign that the company doesn’t expect to build everything on its own.
“Whatever the digital future is, it’s clear that no single company can build it,” said Jeff Butler, chief product officer at Avalon Corp., in the announcement.
The press release notes that the funding for Avalon Corp is led by Bitkraft Ventures, Hashed, Delphi Digital, and Mechanism Capital, with participation from Coinbase Ventures, Yield Guild Games, Merit Circle, Avocado Guild and Morning Star Ventures. Backers also include industry veterans with a visionary view such as Twitch cofounder Kevin Lin, gaming legend Dennis Fong, Charlie Songhurst, previous head of corporate strategy at Microsoft, and Robin Jung, previous CEO of game company Pearl Abyss.
“Connected worlds are evolving at unprecedented velocity, ultimately amounting to a Darwinian game, where the fittest and most useful platforms will thrive and stay alive,” said Jun Park, senior associate at investment company Hashed. “We are excited to back Avalon Corp., led by industry veterans with a visionary view, to help the company realize its potential and pioneer the next wave of interoperable worlds.”
Of course, I’ve always thought that OpenSim was going to pioneer the next wave of interoperable worlds. But maybe they can learn something from OpenSim, about how to move users, content, and messages between different worlds in a fully decentralized way.
Still, it’s nice to see that people are still investing in metaverse projects. For a while there, it looked like everyone was throwing in the towel and rebranding themselves as AI companies.
“This year will be a tipping point for disruptive technology’s role in society,” the report said.
According to the organization, the generative AI technology that was all over the news in 2022 is capable of creating realistic images, videos, and text with just a few sentences of guidance.
“Large language models like GPT-3 and the soon-to-be-released GPT-4 will be able to reliably pass the Turing test,” the report said.
The most famous use of this large language model is in OpenAI’s ChatGPT, but the technology has also been licensed to many other vendors, and Microsoft has already begun adding it to Bing and announced plans to embed it in Office and other Microsoft applications.
The Turing test is an experiment in which a human interacts with another entity via a computer and has to guess whether the entity on the other side is another human or an AI.
Some users are already convinced that ChatGPT is either sentient or is actually manned by an army of humans in the Philippines. And a Google engineer was fired last summer because he became convinced that Google’s version of the technology, LaMDa, had become self-aware.
Now, these tools have become simple enough to use that anyone can harness the power of AI, the report said.
“These advances represent a step-change in AI’s potential to manipulate people and sow political chaos,” the report said. “When barriers to entry for creating content no longer exist, the volume of content rises exponentially, making it impossible for most citizens to reliably distinguish fact from fiction.”
This will have adverse impacts on political discourse. Conspiracy theorists can generate bot armies to support their views, as can dictators.
And companies can also be affected, since key executives can be impersonated by malicious actors, legitimate product reviews drowned in a sea of AI-generated comments, and social media posts can impact stock prices and overwhelm sentiment-driven investment strategies.
Implications for small business owners
If you are a small business owner, it’s time to create a strategy for responding to these threats.
In the OpenSim ecosystem, we’ve occasionally seen instances where individuals were impersonated by someone else in order to harm their reputations — or people created fake personas in order to promote a particular grid or service.
We can expect this kind of activity to accelerate as AI technology allows bad actors to operate on a much more massive scale than before.
At Hypergrid Business, we haven’t — yet — seen a flood of AI-generated comment spam. Hopefully, the Disqus platform we use for comments will be able to filter the worse of it out before we have to deal with it.
Grids that have a social media presence should start thinking about a possible strategy, or a reaction plan, in case something happens. It’s always better to come up with a plan ahead of time instead of reacting in the moment based on emotion, which will usually just make the situation worse.
But there, of course, also opportunities for business to use generative AI for good.
OpenSim grids can use the technology to create AI-powered interactive NPCs to create interesting interactions for visitors to their grids, use ChatGPT to create in-world storylines for users to experience, and use generative art platforms to create textures, images, 3D objects, and even entire scenes.
Grids can also use AI to help create marketing and promotional content such as articles, videos, or podcast episodes.
Marketing is the single biggest challenge that OpenSim grids and service companies have today. If AI can reduce some of the burden, that will be a big win for the whole ecosystem.
In the first decade of this century, unscrupulous website owners noticed that Google had caught on to their tricks — link exchanges, invisible keywords, and meta tags. These tricks were designed to fool search engines into thinking that the websites were better and more useful than they actually were. Spoiler alert: these websites were usually completely useless, just some filler text surrounded by a sea of ads.
After Google updated its algorithms, the scammers found a new trick — content farming. Content farms were companies, often based overseas, that hired armies of poorly paid freelance writers to rewrite existing web posts in slightly different words to create new articles. Then they hired an army of equally poorly-paid editors to clean up the resulting mess into something barely readable.
I get dozens of offers a day from content farms offering guest posts for Hypergrid Business.
I routinely mark them as spam and delete them. In fact, any email that has the words “SEO,” “100 percent unique,” and “Copyscape protected” is pretty much 100 guaranteed to go into the trash bin.
This kind of fake, filler, spammy content proliferated across the Internet. People didn’t usually stay on these pages for long. After all, there was nothing new to read there. But it was long enough for the publishers to get the revenue from the ads on the page and that was all that mattered.
Some legitimate business owners fell into this trap. They thought that in order to get traffic to their websites, they needed to game the system too, and they bought stuff from content farms to fill their pages with meaningless fluff.
For a little bit more money or effort, they could have actually created real content. Something useful to their customers, based on the company’s actual expertise and opinions.
The business owners and website editors that fell into the trap were kicked to the back of the line when search engines caught on — and then they have to work even harder just to get back to where they started.
Content farms are like crash diets. They seem to work at first, but at the end you wind up worse off than you were before, even heavier and lazier, looking for next crash diet, the next quick fix, the next SEO magic wand.
Stop worrying about the scale. Exercise more. Eat healthier food. Okay, I know that doesn’t work for everybody — but crash diets don’t work for anybody at all.
But what I was talking about in that article was taking a bunch of new and useful information and using AI to help you organize it into readable form.
Not everyone is a great copywriter. So if your company has an announcement to make, or a new service it’s offering, or something else they’re doing that benefits people, it can be very useful to have an AI to hammer it into a readable form.
That’s not what I’m talking about when I talk about AI content farms.
(Image by Maria Korolov via Midjourney.)
What I mean is those articles with titles like, “10 top ways to cook with chicken,” except now written by AI instead of desperate writers.
Then you click on the article and find out that it’s a retread of everything already written on the topic, without any new insights or any value. It’s the exactly the same as the old content farm junk, but now written by AI, so it’s cheaper and faster.
These articles are easy to generate. You just ask ChatGPT, “Write me an article about ten top ways to use chicken” and it spits it out.
And everyone out there is doing exactly that, producing the same article in infinite variations.
Now there are companies built around this idea. In case you’re wondering who they are, I’m not going to help you. I’m not going to link to them here and give them free publicity. In fact, the reason I’m writing this article now is that one of those companies contacted me and invited me to their affiliate program and it annoyed me.
I’m not a fan of these companies. Not because they’re charging for something people can get for free from ChatGPT itself. No, they do provide some value. They make it easy and convenient to generate that filler article, and they stuff it full of keywords, and optimize the headline for maximum click bait, and analyze search terms in order to suggest the topics you should cover next.
It’s still the same generic, useless, repetitive content. Just an order of magnitude spammier.
The companies doing this proclaim that their posts can pass AI detection tools. That is completely irrelevant. These articles are still a waste of space.
So don’t fall for the quick fix trap of AI content farms. They won’t help your site in the long run. And, depending on how good Google algorithms are, probably won’t do much in the short term, either. If you fall into the AI content farm trap, you’ll have to fight even harder to get back into search engines’ good graces.
Instead, focus on providing value. Feature your personal experiences, or your company’s expertise and knowledge of the subject. Provide brand-new information that hasn’t been seen anywhere else before. Find a fresh perspective. I did an article about how to do this earlier this month: P.E.A.N.U.T.: 6 steps to staying ahead of AI when writing articles.
So if you’re an OpenSim grid owner or service provider or another company actually doing something and are looking to drive traffic to your website and are considering using AI to help you write some articles — go ahead and use AI. Or you could hire some of those freelance writers that are about to become unemployed.
Just don’t churn out more useless fluff.
And stay away from any vendors that promise “SEO content.” Please. Do us — and yourself — a favor.
And me? I’m going to go add the term “guaranteed to pass AI detection tests” to my spam filter.
I watch a lot of YouTube videos about new AI technologies. It’s a great way to stay on top of what’s happening while, say, cooking dinner or working out. YouTube videos tend to be relatively short and very accessible — a great way to learn about a complicated new topic before you go on to delve into the long-form articles and research papers.
And in my searches for new videos about ChatGPT and Midjourney and other AI tools I kept coming across scammy get-rich-quick videos. The ones that say, “Make $5,000 a day with ChatGPT!” If you’ve ever been on YouTube, you know these guys. They used to pitch crypto and now they pitch AI, but it’s the same dumb thing all over again.
I watched a couple of them to see what they were on about, and the main idea seems to be to post on Fiverr or some other freelance platform offering copywriting or design services. Then whatever the client asks you to do, you just have ChatGPT or Midjourney do for you. A few minutes of cutting-and-pasting and — voila! — the money will just roll in!
The problem with AI-based get-rich-quick schemes
Obviously, this idea is stupid. Why would anyone pay you to do something they can do for free on ChatGPT’s free plan, or one of the many free Midjourney alternatives powered by Stable Diffusion?
Don’t worry, the hucksters say, there are people who don’t know about the free tools, or don’t want to learn how to use them, and would rather pay you money instead.
Umm… maybe…?
Then, the next question is, if all you’re doing is cutting and pasting, then what’s to stop someone from building a simple app that exactly that — cutting-and-pasting from Fiverr customers into ChatGPT and sending back the results? They can flood the Fiverr marketplace with these accounts, and because they’re using automation, they can cut and paste a hundred queries in the time it takes you to do one.
And then they can take some of the money they make doing this and put it towards buying ads, improving good customer service, and so on. And they might even invest some of their profits into building an easy online app that does just that one thing that they’ve automated.
And if they can, they will. I’m sure people are out there doing just that right now as I type this.
That still means you have a window of opportunity if you act fast, the hucksters say. Jump on this trend right now, before the professionals move in, and you can make thousands just by cutting-and-pasting, no skill required!
The problem with this approach is that it takes time to build a business on Fiverr or any other freelancing platform. You have to do a lot of small jobs, for very low pay, in order to get clients who will leave good recommendations for you. Then, once you have built up a reputation, you can start raising your rates and investing in your marketing.
Meanwhile, the people who already have those business aren’t idiots. They’re noticing that ChatGPT is out there, and they’re using it and other tools either to speed up their own production or to improve the quality of their work — or both. And they already have a customer base, and they already know how to write, or how to design, and when they look at the results ChatGPT or Midjourney spits out, they know which ones are good and which ones aren’t, and they know how to tweak them to get them the rest of the way. Plus, they already know how to manage customer relationships and do marketing.
If your only skill is cutting-and-pasting, you’re not going to be able to compete against these professionals.
And that’s the point of most of these videos — you can get rich without having to do any hard work or learning any real skills.
What to do instead
If you want to use AI to get rich, then use it to improve the efficiency or quality of the stuff you are already good at doing, or to help you with the individual tasks that you are bad at. So, if you have a successful business, but you’re bad at writing letters to prospective customers, use ChatGPT to create drafts for you. If you’re bad at coming up with marketing ideas, use ChatGPT for that. If you’re bad at responding to complaints… you get the idea.
Use ChatGPT to build on your positives and reduce your negatives.
You can also use ChatGPT to learn new skills, either to level up in your current profession, or to move to another field.
So, yes, there are plenty of opportunity for people do things with AI. You can use it to help you create content, to create images, to create music and videos. You can use it to create children’s books and comics, YouTube show scripts and almost anything else you can image. But you still have to build a business if you do any of those things. You will still need to find customers or viewers, you will still need to know the difference between good content and bad content, and you will have to figure out how to do things that other people can’t do.
In the virtual world space, for example, you can use AI to generate images, to generate text, and to generate code.
You can use the AI-generated images for textures, or for game maps, or for marketing materials.
You can use text to create games, to create scripts for in-world characters, or for marketing company.
And you can use code for in-world scripts, for server-side applications, and for website plugins.
But, right now, none of these are “set and forget” types of applications. You will have to review the images, text, and code that the AIs generate and carefully select the ones you need. Sometimes you will spend hours, days, or weeks modifying the prompts to get what you want.
AI is not yet ready for real-time embedding into your virtual world because it’s too easy to get a chatbot to go off the rails. Instead, if, say, you want an in-world AI-powered character, use the AI to generate scripts, instead.
Or you could start a company to create AI-powered chatbots that are tightly constrained by the games where they live, can hold a convincing conversations while staying in character, but can’t be diverted by malicious users into spewing racist garbage or crazy conspiracy theories.
When I first tried out ChatGPT late last year, I noticed that many people were using it to write code. Python. Javascript. Even machine language. So of course I asked it if knew LSL, the Linden Scripting Language used in Second Life and OpenSim.
It said no. I was disappointed but not too surprised. After all, LSL isn’t one of the big languages. It probably doesn’t have enough training data.
Today, for some random reason, I decided to try again, and it not only knew LSL, but could explain how it worked. Maybe I worded the question wrong last time?
It can even talk about the differences between LSL and OSSL, the OpenSim Scripting Language, and write OSSL-specific scripts.
Warning: It’s not perfect
We’re still in the beta release of ChatGPT.
It makes mistakes.
So test all code carefully. However, if there’s an error, you can ask ChatGPT to fix it.
And if you don’t understand why it did something, or prefer it took a different approach, you can tell it that, too.
The company behind ChatGPT, OpenAI, is working on improving its accuracy — and the billions just invested by Microsoft will definitely help.
From what I’m hearing, developers are finding that ChatGPT can already significantly speed up their workflow, but that it doesn’t completely replace them. Yet.
When I asked it to recreate the Very Simple Greeting Script by Jester Knox, it took quite a bit of prompting. On the first try, it forgot a key step. When I pointed that out, ChatGPT apologized and rewrote the script with the step added. Then I asked it to use different commands than it had chosen.
All in all, it took it four tries to get to the exact script I was looking for.
My verdict?
Right now, ChatGPT is like a very junior programmer who still makes a lot of mistakes, is supremely confident in their abilities, but is at least willing to admit when they’re wrong.
I can definitely see the potential here for creating simple scripts, especially if I need to tweak those scripts a lot.
But the big deal here, of course, isn’t in what it can do right now, but in what it will be able to do tomorrow.
“Today, the scale of the largest AI computations is doubling every six months, far outpacing Moore’s Law,” Google CEO Sunchar Pichal said in a post last week.
So if a big part of your day-to-day workflow involves writing LSL or OSSL scripts, it might be time to look for ways in which ChatGPT can speed things up. For example, you can ask it to add comments to existing code. You can ask it to outline a coding project. You can ask it to write documentation, or to create slightly different versions of scripts. If you don’t understand how something works, you can ask it to explain it to you.
You can even paste in an existing script — maybe something from the Outworldz free scripts collection — and ask ChatGPT what the script does, then ask it to suggest some ways to use this script in an OpenSim grid.
I get dozens of emails every day from people who want to contribute articles to Hypergrid Business. Sometimes, they even want to pay me to run the articles. The catch? There’s a link somewhere in the article to something the author wants to promote.
I’m not, in principle, opposed to running sponsored content. We’d just put a “sponsored” tag on it and use the money to pay for things that people don’t want to write for free. For example, I’d love to be able to afford to hire a freelancer to go to OpenSim concerts and other events and write about them. Wouldn’t that be a fun job for someone?
But sometimes the article topics are good enough that I’d run them for free. For example, a VR company might have one of their experts write an article about the state of VR in some industry. That could be useful, even if it does promote that VR company.
Unfortunately, 99 percent of the articles that people suggest are worthless.
Here’s how I think they are created:
A content farmer gets an assignment, such as “Benefits of VR in medicine”
They Google the topic
They cut-and-paste the information they get from the Google search
They rewrite the results just enough to pass a plagiarism check and throw in some search engine-friendly keywords
They send me the article
Why would anyone want to read this? Everyone has Google. They can just Google the topic themselves. There’s no new information in this article. It just fills up space and lowers the value of the website to readers. As a rule of thumb, whenever I get an email pitch that doesn’t explain who the author is, that guarantees that the article is “Copyspace-proof,” or that ever uses the word “SEO,” I mark it as spam and hit the delete button. This stuff is garbage and just wastes my time.
Now, AI is going to make the problem even worse because the work flow will be dramatically accelerated:
A content farmer gets an assignment, such as “Benefits of VR in medicine”
They ask ChatGPT to write an article on the topic
They send me the article
Again, ChatGPT is free. Anyone can just ask ChatGPT themselves the same question. There’s no new information in this article that’s produced. Even if search engines don’t decide that all AI-generated content is spam, they’ll still down-rank it because it has no new information.
Google released guidance this Wednesday, in fact, about how they decide whether AI-generated content — or any content, for the matter — will now be ranked by their search engines.
I’ve added a couple more criteria to their list, based on what Hypergrid Business is looking for and created my own acronym — PEANUT — which stands for Personal, Emotional, Authoritative, Novel, Unique, and Trustworthy.
(Graphic by Maria Korolov via Canva and Midjourney.)
The more of these things you have in your article, the higher the chances that I’ll run it. And that applies to both sponsored and non-sponsored posts.
Here are each of those six PEANUT factors, in more depth.
(Image by Maria Korolov via Midjourney.)
P is for Personal experience
The person writing the article should talk about their own history with the topic.
For example, if you’re writing an article about AI in medicine, you might have been treated by a doctor who used AI to help read their scans. Or a relative might have been cured of cancer based in part on a treatment created with the help of AI.
If you’re writing about virtual worlds, you might talk about your own history with the platform.
An AI doesn’t have personal experience with anything, because it’s not a person.
If you don’t have personal experience with a topic, you can get some. You can ask your relatives if any of them have been treated for diseases with the help of AI. Or you might search your memory — did you get a COVID vaccine that was created with the help of AI?
Or, for the virtual world article, you can go and log into a virtual world and try it out.
(Image by Maria Korolov via Midjourney.)
E is for Emotional connection
Why does this topic mean so much to you? Was the relative cured of cancer particularly important to you? Write about that connection.
Did your experience in virtual worlds inspire you in some way? Did it make you feel things you hadn’t felt before? Were you able to do something emotionally meaningful in a virtual world? Write about that.
AIs don’t have emotional connections because they don’t have emotions.
By explaining what the topic means to you, the emotional weight it carries, you help the reader make an emotional connection as well. And that makes for a better article.
A is for Authority
Why are you the one writing on this topic? What makes you an expert?
Really. I want to know. What makes you an expert? Put your resume highlights up top in the story. If you’re the founder of a VR company, say so right up front. Your opinion matters because you know what you’re talking about.
In particular, can you offer some advice or insight that’s better than what an AI can provide, because you have deep expertise?
An AI might recommend a list of things to do, for example, but only you might know what actually works and what doesn’t, based on your experience with customers.
An AI will just repeat all the same advice that’s already out there. It won’t know that some of the advice is worthless or outdated.
(Image by Maria Korolov via Midjourney.)
N is for Novel
Do you have any new information in the article that just came out and isn’t yet available to the general public?
For example, maybe you just went to a conference, talked to a bunch of experts, and learned some cool stuff that most people don’t know yet.
Or maybe you were experimenting with some virtual world tools, and figured out a new hack.
Or maybe you got a press release from a company that is still under embargo, so nobody has seen it yet. If we publish your article right when the embargo lifts, we can be among the first publications that publishes this news.
Or maybe you conducted some research or ran a survey and have new results to share.
If you don’t have anything new to say, find it. You can talk to experts, for example, and see if any of them have a new angle on the topic. Ask them what’s happening that most people don’t know about yet, or about their predictions for what will happen next.
(Image by Maria Korolov via Midjourney.)
U is for Unique
I don’t want to run an article that’s the same as a thousand other articles out there. What makes your article different?
Do you have a different background or point of view? Do you have a particular set of qualifications that makes you uniquely qualified to write about the topic?
When you tried out the technology, did you get different results than everyone else? When you ran the survey, were the results unexpected? When you talked to the expert, did they make a prediction that was surprising, but, when you think about it, makes sense after all?
Or maybe you have an exclusive. That means that something happened and you’re the only one who has this information. Maybe you were the only one at the event when it happened, or a source agreed to only talk to you.
Exclusives are good stuff. Publications love exclusives. Sure, they don’t last long, as everyone else jumps on them soon afterwards, but, for a few days — or a few hours — we’d be the only ones with the news. Score!
(Image by Maria Korolov via Midjourney.)
T is for Trust
Why should we trust your opinion?
If you’re the founder of a company, or the owner of an OpenSim grid, of course you’re going to tell us that your stuff is the best.
If you offer a particular product or service, then of course you’re going to tell us that everyone needs that thing, and it solves all your problems.
Maybe the article you’re writing has nothing to do with selling something. For example, if you’re an OpenSim grid owner and you’ve surveyed your users about, say, how tall their avatars are, then that’s potentially useful information to people who create avatars — and isn’t overly promotional.
If, however, the survey is about why people love your grid, then yeah, we probably won’t run it. I mean, can you imagine that survey? “Do you love our grid more because of the community or because of the great support you get?” Really? Those are the only options? Yeah, nobody’s going to buy that.
You don’t have to have all six PEANUTs for every single article, but the more you have, the higher the odds that I’ll run it.
Let’s look at how this article you’re reading now scores on the PEANUT scale:
Personal: Yup, I’ve got personal experience accepting articles for publication.
Emotional: Well, I am getting a little annoyed about having to go through all those spammy emails.
Authority: Yes, I’m an authority on this subject. I’ve been editing Hypergrid Business since 2009, and, before that, I was a business news bureau chief in China for five years. Not to mention all my other years of experience as journalist and editor.
Novel: As far as I can tell, my PEANUT acronym is completely new in this context — it also means “phase-inverted echo-amplitude detected nutation” in the context of peanut allergies, but I don’t think that counts. Plus, Google’s post was only published on Wednesday, so not many people have weighed in with their advice yet. Google’s own acronym is EEAT and, personally, I think mine is catchier and more comprehensive.
Unique: I am the only person able to comment on how to get stories published in Hypergrid Business because I’m literally the editor of this site, and my word is final.
Trust: Why would I lie about this? I want people to submit good articles so that I have good stuff to run on the site.
Well, look at that. Six for six. I’m hitting the “Publish” button now.
At Hypergrid Business, we’ve been covering OpenSim and other desktop-based virtual environments since 2009, and began covering virtual reality extensively in 2014.
Our goal has been to promote an open-source, distributed metaverse, and to help the little guy and gal with navigating that landscape. Specifically, the people running OpenSim-based virtual worlds — because I think that OpenSim is a great role model of how to create an interconnected, open-source metaverse.
Today, I believe we’ve entered a new era for small business. Well, not today today. More like, on Nov. 30, 2022, when ChatGPT was released.
People worry that AI will take their jobs, and will gut the creative industries. That will probably happen. And there are lawsuits being filed right and left and government regulators trying to figure out how to prevent the worst consequences, and columnists opining about all of this at length. The criticism are fair, and do need to be addressed. I don’t disagree with that.
(Image by Maria Korolov via Midjourney.)
But there’s a flip side to the technology.
AI can be a power multiplier. Like what happens when you replace a stone axe with a power saw. (I’ve been watching a lot of Primitive Technology videos lately.)
In the hands of the big players, it can be devastating. But the big guys tend to be slow to adopt new technologies because they have a lot of institutional inertia and can’t turn on a dime.
In the hands of the little guys, AI can be extremely powerful. Any disruption is an opportunity for a small company to seize market share and grow big, and I believe that AI is going to be a bigger disruption than any we’ve seen before.
(Image by Maria Korolov via Midjourney.)
“Today, the scale of the largest AI computations is doubling every six months, far outpacing Moore’s Law,” Google CEO Sunchar Pichal said in a post on Monday.
Moore’s Law is about technology doubling every two years.
Now imagine it doubling every six months.
Just think of how much change we’ve seen in the past couple of decades — the Internet, the World Wide Web, smartphones. This year, global smartphone penetration is expected to hit 89 percent. Pretty much everyone on the planet, no matter where they live, either has a smartphone or knows someone who does, and can access all of the world’s information. For those without deep pockets, the Android operating system is free, meaning that a couple of years ago I bought my grandmother a smartphone at Walmart for under $30 that did nearly everything my $800 phone did and had a bigger screen to boot — in emerging countries, the phones are even cheaper. Web browsers are free. Web translation is free, for more than 100 languages. Online maps and turn-by-turn directions are free. Weather reports are free. The world’s biggest and most up-to-date encyclopedia is free. You can take online programming courses and earn certificates, for free.
Access to the Internet has put immense power in the hands of individuals and small businesses around the planet. The world is very different from what it was at the turn of the century.
All that knowledge and communication has changed societies — and accelerated scientific progress. We’re editing our own DNA. We’ve cracked fusion and quantum computing. We’re flying drone helicopters on Mars. We’re planning a permanent base on the Moon and building space industry infrastructure.
I am extremely excited to be alive today. It’s scary. All the stuff we’re doing can go wrong very easily, AI included. But there are also immense opportunities.
Sure, not every technology is a slam dunk. There are plenty of platforms out there desperately searching for use cases including, I’m sorry to say, virtual worlds.
AI isn’t in the category. AI is all use cases, all the time.
It’s already being deployed as a force multiplier in every single industry, and is the top priority for pretty much every CEO of every major company on the planet. Until last fall, regular people got to use AI when it was embedded into day-to-day tools, like search engines or recommendation apps. Big companies use AI directly, because they have teams of data scientists on staff.
ChatGPT changed that. On Nov. 30, it made direct access to AI easy and free for people.
ChatGPT grew to 1 million users within five days of its release. In January, it hit 100 million users, according to USB.
“In 20 years following the Internet space, we cannot recall a faster ramp in a consumer Internet app,” USB analysts said in a research note.
According to VR visionary John Carmack, AI will soon be able to simulate the human brain — probably within this decade.
ChatGPT and other generative AI like image generators and voice and video generators, have pushed the boundaries of what AI can do into an area that many experts that I talked to a couple of years ago said was impossible with current technology. AIs with common sense were still decades away, they said.
(Image by Maria Korolov via Midjourney.)
Turns out, common sense and creativity are emergent abilities when AI systems get big enough — some of 137 emergent abilities of large language models like ChatGPT. That means that ChatGPT can do things that can’t be explained by the fact that it is “auto-complete on steroids” or a “stochastic parrot.” Yes, it looks for correlations and uses statistics to predict what to say next. But at a certain point, with enough training data, and enough parameters, weird stuff starts to happen. Like, for example, ChatGPT can pretend to be someone else and answer questions like they would — even if that particular scenario has never been seen before.
That’s the reason why ChatGPT is taking off, while previous AI powered chatbots didn’t. It’s reached a point where it is actually usable. And people are using it. They’re using it for a million minor tasks, like creating recipes out of food they have in their refrigerators, to writing code, to learning complex topics. Yes, it makes mistakes. Yes, it’s bad at math. No, it can’t access the Web. But OpenAI is working on addressing those issues, and so are all of its competitors.
ChatGPT already has a million use cases.
And it’s ridiculously easy to use. Just go to chat.openai.com and ask it questions. But just because it’s easy to use doesn’t mean that it’s easy to use it well. It’s like playing a keyboard. Yes, pressing keys and getting pretty sounds is pretty easy, much easier than, say, trying to do the same on a violin. But getting it to play what you want — well, that’s the trick, isn’t it?
So how do you use AI effectively as a tool? How do you surf the new wave of technology, instead of being drowned by it? That’s the question I’m going to try to answer.
During the day, I write about AI for major tech publications, like CIO magazine. I cover how large companies are taking advantage of AI to improve efficiency, find new business models, and better connect with customers.
At Hypergrid Business, I plan to do the same, but from the viewpoint of individual users and small businesses. Small businesses like OpenSim grids.
AI and OpenSim
So, what that means is that we’re going to be increasing our AI coverage here at Hypergrid Business.
I’ve already been writing about AI on this blog. One of my most popular recent articles is about how to use ChatGPT to write a press release. I used an imaginary OpenSim grid as my example, but I could have used a real OpenSim grid instead. More than a thousand people have read that article, many of whom were hearing about OpenSim for the first time.
Does your OpenSim grid, or OpenSim-related product or service, use AI?
Maybe you have an AI-powered NPC on your grid. Or use AI to help design your website or create content. Do you have an AI-powered game in your world? Do you have AI-powered tech support for your grid or OpenSim hosting company?
(Image by Maria Korolov via Midjourney and Playground AI.)
Let me know, and I’ll write an article about it.
There’s a huge appetite for AI news and use cases out there right now. Tell me about your AI projects and you can get a lot of attention for what you’re working on.
This is also a great opportunity for someone to break into technology journalism. If you’re interested in covering AI — or covering OpenSim, of course — email me. I’ll teach you how to write in Associated Press style, connect you with sources, and help you build your professional portfolio.
Email me at maria@hypergridbusiness.com with story ideas — or to find out how to get started as a technology journalist.
Hi, everyone. This is me, Maria, your friendly local technology reporter and blog editor. You might know me from such things as my annual OpenSim stats presentations at the OpenSimulator Community Conference, my various attempts to run an OpenSim grid on my home computer, or the fact that, since 2009, I’ve written over 2,200 articles about OpenSim and the metaverse — and edited more than 1,000 others — for this very site you’re reading now.
The reason that I started Hypergrid Business was because I could see the future, and the future was the metaverse. An open-source, distributed, decentralized metaverse. You know, like the OpenSim hypergrid. And I thought it was the coolest thing ever, and nobody was writing about it. So I would.
I still believe in the future of an open-source, distributed, decentralized metaverse. And I still think that, of all the technologies out there today, the OpenSim hypergrid still comes the closest to that vision. I believe that the metaverse will change everything. It will change the way we socialize, the way we learn, the way we work. But it might not happen quite as soon as I though it would.
The metaverse is still a technology solution searching for a problem to solve. Mostly, that’s because the technology isn’t quite there yet. The headsets make people nauseous. Zoom is good enough for most virtual meetings. The learning curve for virtual world browsers is too high for casual users. And the killer use cases just aren’t there yet. The only solid non-gaming applications I’ve seen for it so far — the ones worth the money and the learning curve — are when it comes to high-end product design. Really high-end. I’m talking ships and cars and buildings, where distributed teams from all over the world can do virtual tours of 3D mockups and talk about design decisions, safety issues, production hurdles, and marketing plans. These are use cases where a single immersive meeting can save a company tens of thousands of dollars in travel costs and thousands — or even millions — of dollars in prototype costs. Plus, the meetings are typically short enough and don’t usually require too much walking around, meaning that people don’t get quite as sick.
But that’s about it. That feeling of presence — of sharing a location with other people — is magical. But it doesn’t make up for all the shortcomings of either desktop-based virtual worlds, or 3D virtual environments, for most non-gaming applications. So yes, desktop-based immersive virtual worlds — also known as first-person shooters — have taken over when it comes to games. Yes, Minecraft is super popular. But it hasn’t really translated yet to non-gaming applications on any significant scale.
We’re still waiting for killer apps. We might be waiting for a while. I think things will start to change when we have a decent interface or better VR hardare. I know the Firestorm team is awesome, but they’re constrained by having to support Second Life, and the OpenSim community is too small to support its own viewer. I’m worried that Second Life has poisoned the well when it comes to general-purpose, user-build virtual environments because the company did nothing to capitalize on its early media hype.
Interest over time in the search term “Second Life.” (Chart courtesy Google Trends.)
The media — myself included — got excited about the potential of Second Life. It became a plot point in TV shows and made the cover of Businessweek. I guess Linden Lab thought that they’d keep getting free publicity forever, and didn’t bother to invest in any real marketing, user interface design, or anything else that would make it a viable product in the long term.
Second Life’s concurrency statistics have mirrored general interest. Over time, according to data from Grid Survey, concurrency numbers have trended down.
Second Life median daily concurrency, 2009-2023. (Data via Grid Survey.)
The one notable exception to the downward trend was in early 2020, when the pandemic first hit and people were stuck inside and usage jumped for the first time in a decade — but then started trending down again.
Land area, Second Life’s main revenue source, is also down. From a peak of 31,988 regions in 2010, total land area is now down to 27,630 regions, only two-thirds of which are privately owned and a fifth of which is abandoned land.
OpenSim numbers have tended in the opposite direction — both land area and active users have trended up over the years.
Total OpenSim active monthly users. (Hypergrid Business data.)
In the early days of OpenSim, as you can see in the chart above, a substantial portion of the user base was on non-hypergrid worlds, mainly InWorldz and Avination. Those two grids shut down and, since then, nearly all the user growth has been on the hypergrid.
In land area, the difference is even more dramatic, since hypergrid-enabled worlds tend to offer lower-cost land, and some even offer free land, if users are able to run the regions on their home computers.
Total OpenSim land area, in standard region equivalents. (Hypergrid Business data.)
Of course, it’s hard to compare OpenSim directly to Second Life.
First of all, Second Life no longer publishes active monthly user numbers. The most recent numbers are from 2010, when Second Life reported more than 1 million active monthly users. However, historically, Second Life monthly active users have been, on average, 19.8 times that of median daily currency. That translates to as between 700,000 and 850,000 active monthly users this year.
By comparison, OpenSim’s user base is a drop in the bucket.
On the flip side, OpenSim’s land area is more than three times that of Second Life, mostly due to the availability of free and extremely low-cost regions. In fact, the average cost of a standard sized region in OpenSim is less than $13 a month.
What I suspect is happening is that people who don’t need Second Life’s large communities are coming over to OpenSim to get better deals on land and more control over their virtual environments.
You can get a free region on OSgrid by running it on your own computer, or set up a free OpenSim grid of your own using the DreamGrid installer, or have a hosting company set up a grid for you, or rent land from an existing grid. As a region owner, you typically get the ability to save a copy of the entire region in order to keep a backup, or to share or sell to others. As a grid owner, you can restrict who can visit your grid, you can rent out land, and you can run your own currency.
For builders in particular, having an OpenSim grid or region as a must — even if they spend the bulk of their time in Second Life. Building in OpenSim means that you keep the original copies of all your builds in a safe place, and can work with a team on the builds. Building in Second Life is riskier, especially commercial builds created by large teams, due to lack of good backups and ownership issues.
Piracy and copyright aren’t as big a deal. Most content theft happens in Second Life because that’s where most of the content is. And most stolen content is also distributed in Second Life, because that’s where most of the users are. In OpenSim, however, grids have the power to ban griefers and other users who infringe on copyright, and close off hypergrid teleports to grids that don’t follow the law. Plus, the largest platform for content sales — the Kitely Market — has a strong process in place to remove infringing content, with a “report product” button on every individual listing.
The bigger problem is the lack of user base.
For individual users, some find OpenSim grids to be cozier and friendlier than the Second Life environment.
But I suspect that one of the biggest sources of OpenSim users is large role play groups and communities. By coming to OpenSim, these groups are able to set up their own continents at a low cost, or even launch their own grids. They can have more land and more control for substantially less money than they would spend in Second Life.
Another source of OpenSim users is school, which benefit from having closed grids where they can tightly control the content and the people who can access them. However, since these grids are private and generally run behind school firewalls, they do not show up in OpenSim statistics, which are based on public grids.
With the exception of school environments, I personally haven’t met any users who came to OpenSim directly without any experience in Second Life first. That means that OpenSim is fed by the Second Life diaspora, and in intrinsically limited not just by Second Life’s technology but also its dwindling user base.
There’s still a lot of room for growth — we can easily increase OpenSim’s users by a factor of ten or more, just by better marketing OpenSim’s cost advantage.
But, to get beyond that, OpenSim has to able to market itself to the wider world.
As part of that effort, a few years ago, I began covering wider trends in virtual reality. The idea is that people interested in virtual reality are also potentially interested in virtual worlds, and might take a look at OpenSim once they come to the site and find out that it exists. I also offer free ads to everyone in the OpenSim community. Whether you sell stuff on the Kitely Market, have a grid where people can rent land, or run virtual events, or offer building or consulting services, you can get a free ad on Hypergrid Business.
This year, we will be expanding our coverage of AI. There’s a chance that some of the people interested in AI are also interested in virtual environments, and will take a look at OpenSim while they’re here.
(Image by Maria Korolov via Midjourney.)
If you are looking to promote OpenSim to the wider audience, I recommend jumping on the AI bandwagon while interest is high.
For example, you could contribute articles about how to build an AI-powered non-player character in OpenSim. Or how to use AI to generate virtual environments in OpenSim. Or how to use AI to generate scripts.
Also, contact me if you want to become a technology reporter, or columnist, and want to write about AI. Or if you want to write about events happening in OpenSim, or how-to-guides about getting started. I’m also going to be launching a YouTube channel, if anyone wants to be a guest or co-host.
After ChatGPT was released on Nov. 30, 2022, the world changed. Whatever you might personally think about AI, the events of last year showed that AI was capable of human-level creativity in art, music, writing, and coding. And, for the first time, AI demonstrated common sense. Or, at least, something close enough to common sense for all practical purposes.
Companies like Google that had been sitting on their AI projects for years, unwilling to do any damage to their existing business models, are having to rethink their plans.
“The AI race starts today,” said Microsoft CEO Satya Nadella at a press conference today.
The company announced that it’s integrating AI chat into the Bing search engine and its Edge browser — after it invested a reported $10 billion into ChatGPT maker OpenAI last month. The company has also previously announced plans to integrate AI throughout its entire portfolio products.
Adding AI chat to Bing, however, is a direct shot at Google’s bow.
“We can improve the way billions of people use the Internet,” said Yusuf Mehdi, Microsoft’s consumer chief marketing officer, in today’s presentation.
As of the end of 2022, Bing only had a 9 percent share of the search engine market. Google had 85 percent, and the rest was split between Yahoo, Baidu, Yandex, DuckDuckGo and other competitors, all of whom were in the low single digits.
So Bing has a lot of opportunity for improvement.
And speaking of Baidu, a Chinese search engine, it also plans to launch its own AI chatbot, called Ernie Bot. According to CNN, it’s expected to go live in March and is currently being tested internally.
Prices start at $80 per month for a two-region private grid, with $50 setup fee. Includes voice, currency, hypergrid, and a user registration website.
Additional regions are available for $12.50 a month for five regions with a $5 setup fee.
All plans include full backups and root server access. The higher-priced plans have higher limits on the total number of attachable regions, custom domain names, and larger servers.
Add-ons include money servers, region maps, splash pages, in-world search.
Managed hosting also available, for $25 a month, which includes all the regular upkeep and maintenance plus five support requests per month. Additional support requests are available for $25 per five requests.
