Tipping points that alter the course of Internet technology aren’t uncommon, but they don’t happen often either. I think we’ve seen two massive tipping points in 2022: Generative AIwhich jumped to the forefront of internet culture this year, and the beginnings of a movement away from centralized social media towards federated social software. Arguably there was no tipping point at all in 2021, unless you count the rise of WebAssembly. But chances are that generative AI and Federated social software will prove to be truly disruptive technologies for the rest of this decade.
One thing that hasn’t changed is the continued resurgence of web technologies. Last year, it was Figma’s web design tool that impressed me. this year it was web-based development tools – aka “Cloud IDE”. Even with the nascent Metaverse, although Meta wants to make it the next mobile Internet, most of the Metaverse technologies that have stood out this year are web-based.
So let’s go. I’ve drawn from all the articles I’ve written this year (62 of them!), and also reviewed what I’ve been following on social media in 2022, to come up with the five technologies below .
1. Generative AI
If 2021 was the year of Web3, then generative AI was “the new Web3” in 2022. It started with DALL-E 2, which was announced in April and then launched in private beta in July by OpenAI (the company co-founded by Elon Musk, who is certain to be named “Person of the Year” by TIME – it sometimes goes to the bad guys, remember). DALL-E 2 is an image generation service powered by deep learning models, and it was a significant step up from the first version released last year. It was opened for everyone in September and released last month as an API.
The first real competition for DALL-E came in July, when a company called Midjourney launched its eponymous text-to-image generator.
The AI hype really intensified in August, however, with the release of Stable Diffusion – another deep learning text-to-image generator. Unlike DALL-E 2 and Midjourney, Stable Diffusion has a permissive licensing structure. The company behind it all, Stability AI, declared that it was “released under a Creative ML OpenRAIL-M License” – not fully open source, but it allows both commercial and non-commercial use.
We’ve seen a number of variations on generative AI this year, including text to video (Google’s Runway and Imagen Video and Phenaki), 3D text (Nvidia’s GET3D and Google’s 3DiM), speech to text (OpenAI’s Whisper), and of course text-to-code (OpenAI’s Codex, which powers the GitHub co-driver). Although Copilot was launched last year, Microsoft announced this year that it has become an important part of its developer tools. Microsoft, which is a major investor in OpenAI, told this year’s Build conference that GitHub Copilot “today suggests about 35% of the code in popular languages like Java and Python”.
Last but not least, we got OpenAI ChatGPTreleased this month, which is a chatbot for generative AI.
2. Mastodon and the Fediverse
After the the news has come In late April, after Elon Musk’s offer to acquire Twitter had been accepted, many Twitter users began to rekindle their old Mastodon, Tumblr and other social media accounts. But at the end of the year, there was only one alternative for Twitter: Mastodon.
When Musk finally took over Twitter in late October, after months of trying to wriggle out of the deal, he immediately did drastic changes. This led to a full migration to Mastodon, which continues to this day. The open protocol that underpins Mastodon, ActivityAdallows it to “federate” with completely different social networks, like Pixelfed (a photo-sharing site) and PeerTube (a YouTube clone). This new open web platform has been dubbed “the Fediverse”.
What’s fascinating about this is that fedivers represent a return to decentralized social networking, not unlike the first wave of blogging in the early 2000s. Switching away from centralized platforms will ultimately spell the end of Twitter, it made decentralized software again attractive for developers. In the month since Musk took over, Twitter’s dev platform team has been decimated (ironically, I had wrote about it the previous month, cautiously optimistic that Twitter was back on track with developers). Simply put: there’s nothing left for developers on Twitter, which leaves a blank slate for the fedivers. Watch this space in 2023!
3. Cloud IDEs
In September, the internet industry’s deal of the year happened – Figma, a popular browser-based design tool that I selected as one of my top technologies of 2021, sold to Adobe for about $20 billion. After that, attention may have shifted to the last category of productivity tools that has yet to hit the web: integrated development environments (IDEs).
Later that month, Ivan Burazin, co-founder of a web-based IDE product called Codeanywhere, told me that almost all external developers soon to use a Cloud IDE. Internal developers will also increasingly have the option, although this is only a feature of enterprise tools like Microsoft Visual Studio.
I also spoke to Shawn Wang, aka @swyx, Shawn Wang, who told me we were about ten years away from devs completely delete desktop files – it will be “the end of localhost”, he said.
If this trend interests you, discover my interviews with the founders of StackBlitz – “quite the development analogy of what Figma did for design”, according to StackBlitz co-founder Eric Simons – and code sandbox.
4. Open the Metaverse
Mark Zuckerberg’s company continued to dominate the discourse on the so-called ‘metaverse’ this year, following its giant pivot and renaming to Meta in October 2021. However, at the end of 2022, there were still few of evidence of what the Meta platform will actually be built with.
On the other hand, it was a year of solid progress in the web-based open metaverse – including projects like the Open Metaverse Interoperability Group (OMI), the Metaverse Standards Forum (a Khronos Group initiative) and various Web3-branded virtual worlds (Decentraland, Voxels, Webaverse and others).
What’s perhaps most interesting about the open metaverse is how it’s changing online culture. In June, I spoke with Jin, one of the leading open metaverse developers. We met in a 3D virtual space in Mozilla Hubs, as Jin is a pseudonym, meaning he hasn’t revealed his true identity online. This, it turns out, is an important feature of the open metaverse. According to Jin, “pseudonymous work will be important in the virtual economy, avatars are inherently a privacy-preserving technology.” The pseudonym also reflects key influences on Jin and others like him, which include gaming culture, the first wave of online VR headsets (led by Oculus), cryptocurrencies, and “Vtubers” (short for ” Virtual YouTubers”, online artists who use avatars).
A number of fascinating new startups and metaverse projects have sprung up in 2022, including Croquet and Third bedroom. Nvidia’s Omniverse, which I picked as one of my top products last year, also made further progress with his tools. While Meta has done some encouraging sounds in October – particularly with the announcement of its new Meta Quest Pro VR headset – 2022 proved that the open metaverse will be a viable alternative to Zuckerberg’s vision in the future.
5. Decentralized storage
Web3 was arguably the biggest tech story of the last year; or as I said my annual overview, “a ball of hype that has gathered momentum in 2021.” However, in 2022 things went downhill due to a not hard to predict cryptocurrency crash. But even though the development of blockchain technologies slowed down as a result, there were still pockets of innovation in Web3. One was the continued advancements being made in decentralized storage.
Probably the most impressive of these solutions is the Interplanetary File System (IPFS). As fellow The New Stack writer Jake Ludington noted in september, while normally IPFS is used to host media resources for blockchain applications and NFTs, its usefulness extends beyond Web3 implementations. “The decentralization built into IPFS is available to all application development and has the potential to change the way web applications are built and scaled,” Ludington wrote. He suggested developers consider using IPFS instead of sticking with the HTTP stack.
Another interesting project is the almost as aptly named internet computer (ICP), which sits between a blockchain and a decentralized cloud provider. It is not a storage solution per se, as it is made up of independent data centers that cooperate with each other (like a peer-to-peer file sharing network). According to an interview I conducted with the organization that runs ICP, developers can host an entire application there, which, it is claimed, obviates the need for a traditional cloud service like AWS.
Arweave and Filecoin are two other solutions to watch out for. Even the Internet Archive, that modern web treasure, is actively researching decentralized storage solutions.
This top 5 list is of course entirely subjective, though I promise I haven’t reviewed ChatGPT or any other generative AI. If you have any other suggestions for the best internet technologies of the year, I’d love to hear them — tag me in fedivers to start a discussion (oh ok, Twitter if you must).
To note: the signature image for this post was generated by Stable Diffusion, with the prompt “generative AI is taking over the internet, but people are turning to the juggernaut”.