You hear a lot about chatGPT and GPT-3 these days. For some it’s the birth of a real AI, whilst other feels it’s just an overhyped gimmick. My take is that whilst it is definitely not “real” AI (whatever that means), it’s a very valuable tool that is here to stay. It will transform the way most of us access information. But its inhate flaws (hallucinations, bias) as well as its potential for abuse (fake news, sponsors) will make us regret the days of getting our information from Facebook.
In the following piece, I’m talking about Chat GPT as a system to access information. It has many other uses (creative writing, language learning, code transformation etc.) but I’ll focus on accessing information here.
1. Web 3.0
The Web 3.0 is an expression I’ve heard used left and right since I started my career. Often used to described some gimmick features powered by JavaScript as a marketing ploy. Most recently, it has been baffling to hear the excitement around the Metaverse and NFTs as “the Web 3.0”. A stupider version of second life and a pyramid scheme are quite far to my definition of Web 3.0. (Why I think NFTs, Cryptos and the Metaverse are not the future is another topic altogether, perhaps for another post.)
But what would a definition of web 3.0 be?
Well the Web 1.0 was about giving static information. You would go on page created by Webmasters (Or, fun fact, sometimes called “Webmestre” in French, which I just imagine as a guy in a robe with a staff and a big hat). You would read their content and that’s it. If you had something to say yourself, you would create a website of your own.
The Web 2.0 changed the way we related to that information by making it interactive. Forums, Comments Sections, votes, etc. That was a massive change in how we interract with information. Then came social media, Facebook, Twitter etc. At some point, a major shift started to operate: What information you saw was chosen by some proprietary algorithm, with interests of its own. This led infamously to the rise of fake news and the echo chamber. The algorithm did not care about presenting you with the right information, it cared about keeping you on the platform as long as possible. It was the birth of the “attention economy”.
So where does that live us for the Web 3.0? I believe to deserve such a name, it needs to be a major shift in how we access information. Sure, an AR system that is accessible by a high percentage of the population, light and worn all the time, à la Black Mirror, would work there. But I don’t think we’re anywhere close to that. I’d add that the Web 3.0 does not have to be “better” by any sort of moral assessment, it just needs to replace the Web 2.0 as our way to access information. So what could be the next shift in how we access information?
2. ChatGPT
If you’ve not heard of chatGPT, I have no idea under what rock you’re living. I won’t bore you with the detail of how it works as I assume anyone who furiously clicked on this article already has quite a understanding of the topic. With that said, I have found myself quite attached to that new interface to access information.
Whereas googling requires
– Thinking of a query that is generic enough to yield results
– Filtering through pages of SEO generated content, often with a lot of ads
– Reading the content that will usually dilute the information you’re looking for in the middle of a giant wall of text, again for SEO purposes.
Searching on chat GPT means I can make a query specific to my situation.
I hear you. “The information can be incorrect”. It happens. Recently I searched for an Onsen closed to the city of Zao in Japan and ChatGPT suggested one 5 hours away, telling me it was a walking distance.
But honestly, just taking a grain of salt with the information given, and potentially verifying it with a subsequent search seems like an efficient way to get information. One might say good enough.
3. The regular web is going to get worse
Whereas you might think “well, if it’s not broken don’t fix it”, I believe the web as a way to access information is getting worse by the day. Content generated with GPT-3 is going to start to show up for every long tail search under the sun, whereas regular content is going to get even heavier with SEO keyword to survive. The web is going to get worse and worse, and the only way to get good information is with a system that can extract the signal from the noise, a.k.a ChatGPT.
4. Adtech
But the final nail in the coffin of the regular web is the adtech industry. The web is made to show advertisement, except for a few paywalled newspapers and the magical wikipedia, anything else is here to shove advertisment down our throat. Tiktok and facebook have proven to be horible ways to show relevant advertisement that people will click on. But I believe ChatGPT is going to outshine them all.
Imagine a world where sponsors pay good money for ChatGPT to recommend their product. Depending on the ethics of OpenAI (and their competitors), those sponsors could be behind disclaimers or straight as part of GPT answer. Imagine the following conversation
Me: “I’m looking for a nice onsen town near Zao”
ChatGPT: “Well you could try the one in Naruko. It’s a bit far but it’s a nice onsen town. Make sure you own a Yukata before going there though, like this one http://sponsoredlink.com.
Me: No thanks, I already own a Yukata.
ChatGPT: oh alright then, let me know if you need anything else.
Here the sponsor is obviously the Yukata shop. Or is it? Maybe the town of Naruko paid money to OpenAI to be recommended first when people look for onsens. Also with the chat feature, I can tell chatGPT what I’m not interested in, which allows it to get more information about me, which is always valuable. Go block this with an adblocker. Adding a “cuteness” layer to the chatbot could make that even more appealing to teens. And we’re not just talking about selling products here, as Facebook shown, this can be used for political purposes as well, which is all the more dangerous. It can obviously be more subtle than this previous example, slowly changing the behaviour or belief system of its user over time, much like an electronic version of Tom Riddle’s diary.
I can only imagine there are a few dozens stealth startup working on exactly that at the moment.
5. So is there any hope
I don’t think so. Sorry for being so grim, but I believe we are at the dawn of a brand new way to access information. Systems like ChatGPT are going to become more prevalent, with new features such as
– Retaining information about the user
– Double checking the information with other sources
– Presenting actual results for the web (which allows for links to be clicked on, aka money to be made)
– Personalities (the French have the hilarious “Chat CGT”, named after a union, which answers every question with the view of a overly Marxist person, but I would imagine in the future all shades of worldviews could be represented.)
As the current web becomes worse, and because humans are naturally lazy, this will become the new way to access information, at least for the generation of kids currently in school. The fact that that information is sometimes wrong will just be accepted as the price to pay for the convenience. Similarly to how wikipedia replaced encyclopedias even though “anyone can edit it”. My generation will tell their kids “Hey, be carefull of generated content” much like our own parents told us “Don’t believe everything you read on the internet”, before probably falling pray to it ourselves. The best I can hope for is that some hacker collective manages to make an open source version of it, that can be more trustworthy than the current one. But I’m not holding my breath. Maybe a legislation could help, though I have trouble imagine what law could be passed to prevent this.
Bonus
As I was writing this article in VSCode, Github Copilot suggested the following to me:
and I for one, welcome our bot overlords.
Leave A Comment