The One with a Fast Car to DehumanAIzation #2
# intro
Hey, 👋
January has come and gone, along with the first week of February. Time flies, especially when you spend it on one thing and one thing only - filling out application forms and writing cover letters. The job search is still in progress, and I hope to have more updates on it soon.
I've made up my mind about the structure of this newsletter. If you've ever used Slack, you'll feel right at home. If you haven't, it will be easy to grasp the idea and scan through it every week.
Happy reading! 📖
# general
Apple has started selling Apple Vision Pro in the US, so my feed is filled with first impressions. I follow a lot of tech people and can't mute everything, so it feels like a self-inflicted curse on my part. Alas, since I'm poor and not in the States, I spent my week reading other people's thoughts about and contemplating the latest announcement from the Browser Company of New York (the makers of the Arc browser).
I've been using and advocating for Arc for almost 2 years now. I love it, and there are so many things they get right. Unfortunately, it seems they're now doubling down on AI, which is not the direction I hoped for - the latest LLM and AI features miss the mark, at least for me.
The scientists at the Browser Co. were so preoccupied with whether they could go on a crusade against Google, they didn't stop to think if they should. Arc is probably working on their own engine, but I find it pretty ironic they go against Google while using Chromium under the hood.
I’m not saying Google is good, they have been at the forefront of internet enshittification for some years now (for example, they retired the cache link this week and I genuinely believe that this is a shitty move, which makes the internet worse).
However, all the features shown in the Act II video and Arc Search app are not good moves either. They deprive website owners of the motivation to create content and share it online, at least without some kind of "paywall" before it. Moreover, we'll see more and more websites disallow ChatGPT or any other LLM crawler bots from now on. Vox Media is already doing that, as far as I know. You can find a good breakdown of this situation in this thread from Gergely Orosz. Josh Miller, the CEO of Browser Co., addressed the criticism on Twitter.
Right now, they're pouring resources into integrating AI functionalities, tapping into ChatGPT, Anthropic, and various other APIs to enhance their browser. Yet, this significant financial commitment might just fuel the ongoing dehumanization of the internet. While Google's methods aren't flawless, they somehow manage to offer a more appealing alternative. Browser Co.'s mantra appears to be, "Let's disrupt the status quo and take Google's place. We'll sort out the details later." I find this stance worrisome, especially coming from a company without a solid plan for making money. And I'm not even talking about the creators here.
In addition to all of the above, there's also this: Arc Search's AI responses launched as an unfettered experience with no guardrails:
But what stood out most of all during my testing period was that this app had no apparent guardrails in place, and would do its best to give a straightforward answer to — as far as I could tell — literally any question, with sometimes deeply disturbing results.
Browser Co. isn't the only one; Perplexity and Poe are similarly leveraging AI in their search functionalities. On the other side, consider the Kagi Small Web initiative as a contrasting approach that tries to bring a more human touch to the web. Kagi's search engine came highly recommended in an Installer newsletter, and their Orion browser was personally suggested to me on Threads. Though Orion may not be as polished, I'm intrigued to explore both Kagi's search and browser offerings. Even if I have to pay for it.
The web's current state is a far cry from perfect, and the ideal path forward for the internet we cherish remains unclear. ActivityHub? RSS? AT Protocol? Fediverse? Or dare I say, web3? I don't know.
What I'm certain of, though, is that exploiting authors' content without proper acknowledgment, and deterring content creators from launching new websites and monetizing their work, is definitely not the solution.
# privacy
Tech CEOs were dragged before Congress again to answer questions about child safety on their platforms. Mark Zuckerberg delivered an impromptu apology, and we also found out that Singapore is not China. As someone recently mentioned on Threads, we are in a new era - senators and lawmakers now also aim to score clout points during such hearings, using snippets of the proceedings to "build their personal brand" on social media (example).
That note hits the nail on the head, and here is a story about it on Vox Media - Everyone’s a sellout now. It's no longer enough to be a good artist, writer, or any other professional today. You have to "build a brand" and "have enough following", and it's incredibly draining. Prefer to stay private and make a living off your craft? Bad luck for you then.
Speaking of dehumanizing. While researchers publish a study on how computer vision papers objectify human beings and dehumanize people, Moscow has been living in the future for a while already. This article, Inside Safe City, Moscow’s AI Surveillance Dystopia, is one year old, but I don't think the situation has gotten any better. NtechLab, the company behind the system, is under European sanctions from July 21, 2021.
Here is another example of such dehumanization - this coffee shop uses AI and computer vision to measure the productivity of their employees and the time spent in the shop per customer (Reddit, Threads). Given the total disregard of human integrity and privacy, I'm sure it's Russia too. There are other clues as well:
Some people say it's fake, but I don't think it's that impossible to spin up something like that in your cafe.
Last week, Russian Twitter was abuzz with the story of a guy who trained a bot to select Tinder profiles according to his preferences. He then used ChatGPT to initiate conversations and arrange dates on his behalf. He interacted with over 5,200 women, and following ChatGPT's advice to propose, he's now engaged to one of them a year later. There's no way all of this was done by one person, so the story is most likely fabricated. However, I have no doubt it'll happen someday, so we should all be prepared for such stories in advance.
# links-of-note
The Electronic Frontier Foundation suggests a simple yet clever way to combat voice call scams - creating a family password (EFF article). I also wonder if such password would have helped this guy who paid out $25 million after video call with deepfake CFO.
In this interview, Drupal creator Dries Buytaert says that websites are now needed more than ever (read here).
A good post from Anil Dash about how the old technology behind podcasts might mean more for the future of the open web (read here).
A shameless promotion of my own post about the dying internet. Check it out here: Long Live the Internet. And don't miss the previous issue of this newsletter either: The One with AI, Crusade, and a Soviet Cartoon.
# water-cooler
I can't pinpoint exactly when "Fast Car" made its way into my playlist. It certainly wasn't after Tracy Chapman's iconic performance at Wembley Stadium in 1988 (watch here), as that was a month before I was born. But I'm truly glad it did.
Last week at the Grammys, Tracy Chapman shared the stage with Luke Combs to perform “Fast Car” and I must say, her voice is still incredible. Oddly enough, the full performance isn't on YouTube, but you can catch it here: Tracy Chapman and Luke Combs - Fast Car.
Seeing how happy people were to see Tracy on stage after her hiatus made me realize how much we all need more human connection in our lives, both online and offline, not less.
---
Until next week,🖖
Greg