🤖 The AI hysteria is real
No robot uprising for now.
👋 Welcome to FWIW by David Tvrdon, your weekly tech, media & audio digest.
In this edition
There was panic about an AI gone sentient
Elon Musk talked to Twitter employees
More than 1.5 billion people watch YouTube Shorts
Note for 🇸🇰 readers: In Slovakia, I have co-launched a new portal podcast database with all Slovak podcasts out there to help with discovery and also to bring podcast news and information to local community. You can find it here: skpodcasty.sk. For podcast creators we have a fun feature - Spotify and Apple Podcasts podcast charts.
A crazy week for AI conspiracies
A week ago, The Washington Post published a story of Google engineer Blake Lemoine who worked for Google’s Responsible AI organization. It details how he started having conversations with LaMDA, Google’s artificially intelligent chatbot generator, and began to be convinced the AI is sentient.
After he sounded the alarms and nobody did anything, he leaked the conversation with the AI publicly.
For a few days, some believed (or at least the news headlines made it sound that way) that Google is hiding a sentient AI and a new era has begun.
First, I came across some snippets from the conversation and was impressed.
Then, I started reading about it more and became suspicious.
Fortunately, scientist and author Gary Marcus wrote a rebutal:
Nonsense. Neither LaMDA nor any of its cousins (GPT-3) are remotely intelligent. All they do is match patterns, draw from massive statistical databases of human language. The patterns might be cool, but language these systems utter doesn’t actually mean anything at all. And it sure as hell doesn’t mean that these systems are sentient.
Which doesn’t mean that human beings can’t be taken in. In our book Rebooting AI, Ernie Davis and I called this human tendency to be suckered by The Gullibility Gap — a pernicious, modern version of pareidolia, the anthromorphic bias that allows humans to see Mother Theresa in an image of a cinnamon bun.
Indeed, someone well-known at Google, Blake LeMoine, originally charged with studying how “safe” the system is, appears to have fallen in love with LaMDA, as if it were a family member or a colleague. (Newsflash: it’s not; it’s a spreadsheet for words.)
It’s almost striking how simple it is to unpack once you know the underlying biases and understand AI a little better.
Also, Lemoine cut up the conversation so that it appears to have more sense than it might have originally had if he left all the parts.
I encourage you to read the whole blog post by Marcus. It’s really insightful and might make you immune against future bombastic news like this.
Still, it’s fascinating if you think about it, how good the current language model predictions are and I guess we can’t even imagine how a real sentient AI might react.
TECH
😶 Elon Musk told employees he wants Twitter to be more like WeChat and TikTok. In his first Q&A with Twitter employees, Musk shared his vision for Twitter to reach 1 billion users. During the call, Musk reiterated that he wants Twitter to lean more heavily on subscriptions and payments, suggesting that people should pay to be verified. Also, he confirmed layoffs will coming. [The Verge]
And of course, the transcript of the whole Q&A leaked, if you want to read it.
🤨 The UK competition watchdog is considering launching an investigation into Apple and Google’s dominance of the mobile browser market. The Competition and Markets Authority (CMA) indicated that action was needed, saying that otherwise the companies were likely to strengthen their grip on the sector, which would further restrict rivals and dissuade innovators. [The Guardian]
👏 The European Commission released an overhauled set of rules designed to stem the flow of disinformation. The EU's strengthened Code of Practice on Disinformation will hold signatories to the code liable for failing to take action by fining them up to 6% of their global revenue. Tech giants Meta, Google, TikTok and Twitter are among the signatories. The revised rules cover "manipulative behaviors," including deepfake videos, bots and fake accounts, but also aim to eliminate financial incentives for the spread of disinformation by ensuring that disseminators of problematic content don't benefit from advertising revenue. [CNET]
🥱 Samsung has a Wallet app as well. The step is more like the company is creating a new app as an umbrella for all the existing services it already had. It is available today in the US, UK, and parts of Europe for Galaxy phones, with Samsung Pay running Android 9 or later. [The Verge]
😵💫 Facebook employees were recently given a new directive with sweeping implications: make the app’s feed more like TikTok. Facebook is going to change the app in many ways - Stories & Reels at the top of the main tab, followed by posts from its discovery engine (so, no friends and news I guess); bringing Messenger back into the blue app; and more visual, video-heavy content featured. [The Verge]
😲 Nothing reveals Phone 1 design a month early. It’ll be officially announced on July 12th. [The Verge]
🤔 The shift from proof-of-work to proof-of-stake for Ethereum is happening and miners are not happy about it. All that equipment for nothing. [Bloomberg]
🤓 Adobe plans to make Photoshop on the web free to everyone. For now, it’s testing a freemium version in Canada. [The Verge]
🧔🏻♂️ Father’s Day is coming and here are three lists of things I want and will buy myself to post in your family chat as a reminder: Wired, Strategist, The Verge.
MEDIA
⚽️ Apple got the rights to stream Major League Soccer for 10 years. It should be available everywhere on Apple TV. [TechCrunch]
😮 If you were wondering, TikTok hasn’t seen an advertising slowdown despite other social networks having this problem. Blake Chandlee, TikTok’s president of global business solutions, gave CNBC an interview and it’s quite interesting. He said they are not a social network like Facebook, but an entertainment platform. [CNBC]
🎥 More than 1.5 billion people watch YouTube Shorts every month. That’s huge and it’s nice to see Google disclose the number. [WSJ]
👀 An interview with the new NY Times editor-in-chief. [Variety]
😎 A re-examination of Kevin Kelly’s famous proposal that an artist (or creator) could make a living online with a thousand true fans. So, is it true? TL;DR remains to be seen. [New Yorker]
📊 The world’s top ten non-English digital news subscriptions. No.1 Chiense Caixin, No.2 Japanese Nikkei, No.3 Garmen Bild. [Press Gazette]
FROM THE FIX
2️⃣ What Analytics taught me about my community
3️⃣ Your newsroom probably needs a YouTube specialist
4️⃣ Digital audio revenues up 500%, reports AOP and DeloittePress Prize
[ 📬 Get The Fix newsletter delivered to your inbox every week with the latest insights, news, and analysis about the European media market. Sign up here > ]
AUDIO
📚 Daniel Ek says Spotify's upcoming audiobooks vertical will adopt the freemium model. Just a note: The audiobook business has stayed roughly 3x bigger than podcasting for the past few years. [The Verge]
🎙 Substack published a guide to going paid with your podcast on their platform. [Substack]
🎥 Apple added a new page describing video podcasts on the platform, how to publish them and all the features. [Apple Podcasts]
🇫🇷 The most popular podcast in France (Transfert) had 818,000 monthly download in the country an 1 million globally. [Podnews]
🔊 The Smart Audio Report found 62% of Americans (18+) use a voice assistant and 35% now own a smart speaker. [NPM]
❓ Poll: Are you afraid of an AI becoming sentient?
🙌 Thanks. I used HandyPolls to create this poll (instructions).
Last poll results: What are you the most excited about that was announced at this year’s WWDC? 32% iOS 16 features, 12% equally excited for MacBook Air M2 + watchOS 9 + Continuity Camera.
🙏 And big thanks to Celine Bijleveld who helped me edit this newsletter. You can follow her on Substack here.