Convert anything into a podcast
After reading Evan Hatch’s famous description of Len Sassaman (who Polymarket briefly mentioned was most likely Satoshi Nakamoto), Thrive Protocol’s Ben West uploaded the article to Google’s NotebookLM audio compendium.
We quickly put together a 12-minute podcast with two hosts who crack jokes while revealing Len Sassaman’s story.
I played the results to my girlfriend and tested her reaction before revealing that they were entirely AI-generated. She says she was completely fooled.
“It totally sounds like people talking nonsense on podcasts,” she says.
Others were also impressed by the audio outline, which allows users to upload website or YouTube links, PDFs, plain text, Google Docs, and slides.
The audio overview then provides in-depth information about the content in an interactive podcast format. Two natural voices interrupting each other and laughing at each other’s lame jokes.
A16z partner Olivia Moore uploaded 200 pages of court documents to the service.
“It creates a true crime podcast that’s better than 90% of what’s out there right now, and it ends with the hosts debating the ethics of the genre.”
Business Insider’s Ana Altchek also tested the software and said the voice quality was higher than anything else on the market, including ChatGPT’s advanced voice mode.
“Not only do the voices sound real, but the way the AI hosts bounce off each other and joke around mimics the style and structure of a real podcast.”
Although she found no obvious inaccuracies in the results, she certainly has concerns about anything generated by AI. So it may be the best way to get the gist of a lot of complex information in a compressed burst, but you should double-check your facts before relying on it.
“It’s rare to be surprised by the tools you can actually play with these days.”
“Audio Overview is probably my favorite AI tool I’ve ever used, and it shows innovative ways we can use AI in our everyday lives.”
The AI+ cryptocurrency has surged more than twice as much as Memecoin in the past 30 days.
According to Artemis, the AI segment has grown by a weighted average of 38% over the past month, outperforming its closest competitors. Bridge Token rose 38.3% and Mimcoin rose 20.8%.
Among the large AI projects using tokens, Bittensor (TAO) appears to be doing most of the heavy lifting, up 120% in 30 days. Other popular projects also fared poorly, including Arkham (ARKM) up 31%, Artificial Superintelligence Alliance (FET) up 19%, and Near Protocol (NEAR) up 22.2%.
However, over the past 12 months, the situation has turned around, with the AI + cryptocurrency sector down 14.4%, while Mimcoin is up 57.2%.
So the popular AI + cryptocurrency tokens today are probably similar to those from a year ago. On the other hand, holding most memecoins for more than a year is risky for your wealth.
Both centralized exchanges (77.8%) and real assets (351.4%) surpassed Mimcoin despite the term ‘Memcoin Supercycle’, with Bitcoin at 40.6% and Ethereum at 40.6%… *cough* 3.8%.
Also read: Advanced AI Systems Are Already ‘Self-Aware’ — ASI Alliance Founder
true fake
So far, it seems like AI misinformation and distorting the truth is a big problem, but two recent incidents suggest that the bigger problem is getting someone to care.
Some people don’t seem to make that much of a fuss even if it’s proven to be an AI-generated deepfake, as long as they “feel” it’s authentic.
Utah Senator Mike Lee posted a photo of a crying girl wearing a life jacket and holding a sobbing dog on a boat during floodwaters during Hurricane Helene. He captioned this photo so his followers could vent their frustration over the Biden administration’s failure to stop the crying girl and puppy floating around.
Lee deleted the photo after users pointed out it was fake, but Trump stan and groupie Laura Loomer reposted it. “It doesn’t matter where this photo is,” said Amy Kremer, co-founder of Women For Trump. “Photo source,” he said. “There are people who are going through much worse things than what is shown in this photo. So I’m leaving.”
On the other side of the political fence, BBC disinformation reporter Marianna Spring traveled to Pikesville, Maryland in January to follow up on the “racist AI deepfake” scandal. In the scandal, a local school principal was apparently caught on tape making racist comments.
The tape was quickly exposed as an AI deepfake by the Baltimore Banner, but many readers did not believe the explanation, seeing it as a way to avoid responsibility.
Also read
characteristic
Bitcoin Goes Physical: Art or Digital Heresy?
characteristic
‘Make sure Ethereum wins’ — Steve Newcomb reveals key guidelines for zkSync
Six months after a disgruntled school PE teacher was charged with making counterfeit items, many local residents still seem to think the case is real. Spring described a five-minute conversation with a local named Sharon. She continued to talk about the video as if it were real, but her husband gently reminded her that it was fake.
“She admitted that she later found out it was AI-generated,” Spring said. “But he said he was still angry.”
Spring said he’s found time and time again that AI fakers have a hard time getting a first impression in their heads when they’re telling the truth about what viewers already believe.
“Well, even if it’s not real, I guess that’s what they think.” Spring talks about the thought process.
Google searches return a higher percentage of AI images.
If you search for a photo of a baby peacock using Google Images, you only have about a 25% chance of finding an actual photo. X user Dinesh posted a screenshot of the top 15 results for ‘baby peacock’ and said 11 of the images were AI-generated.
“Most of the images were generated by AI. The Internet of Man is coming to an end. The dead internet theory will become reality in a few years.”
The dead internet theory is a conspiracy theory that the internet is mostly fake and not man-made.
Google plans to add disclaimers to AI-generated images in the coming months, but only for images that contain C2PA metadata.
Notable signs of the AI generation include extra fingers, reflections in human eyes, the fact that everything is too perfect, and backgrounds are rendered less well. But as technology advances, these signs will disappear, so platforms must address the issue now.
Earlier this year, Fox Corporation partnered with blockchain platform Verify to track and authenticate the provenance of articles and multimedia, with Time magazine becoming its first customer.
Advertising becomes more cunning when AI takes over
TikTok has officially launched its “Smart+” AI-based ad buying and campaign management tool to compete with similar products from Google’s Performance Max and Meta’s Advantage+.
TikTok has never been the most profitable advertising platform. It brings in $22.32 billion a year compared to Meta’s $154 billion, and it’s hoped the new platform will help change that. The new suite allows marketers to decide whether to create and optimize ads and placements or hand them over to black-box algorithms.
Inevitably, they will hand it over to Algo. Because it is much better at manipulating human behavior than advertisers or marketers.
“Advertisers who optimize for value through Smart+ web campaigns see, on average, a 53% increase in return on ad spend,” said Adolfo Fernandez, head of global product strategy at TikTok.
Ray-Ban saw a 50% reduction in cost per acquisition and a 47% increase in conversion rates after using the platform in beta. Overall, the sunglasses manufacturer saw a 42% increase in return on investment.
Also read
characteristic
Can cryptocurrency exchanges be trusted after the FTX collapse?
characteristic
11 key moments that made Ethereum the No. 2 blockchain
Google’s AI Overview Sells You the Stuff
Google’s AI Overview feature shows users small snippets of AI-generated answers to their questions. At the beginning of his debut, he was criticized for telling people to eat rocks or teaching them how to glue cheese to pizza, but he seems to have improved.
But decades of efforts to make Google searches worse with sponsored ads and SEO spam results mean the company is now serving ads with AI outlines.
If you search for “how to remove grass stains from jeans,” AI will give you some tips, including using a stain remover. Then you’ll see an ad for the helpful TidePen stain remover.
Google makes money from these ads, but it won’t share the revenue with publications and websites whose AI scrapes the answers, the search giant told Bloomberg.
Google’s share of the $300 billion annual search advertising market is expected to fall below 50% next year, while rivals such as Amazon are expected to grow twice as fast, taking up 22.3% of the market this year.
AI search startup Perplexity has been evaluated as having excellent search results, but plans to introduce advertising later this month rather than relying solely on subscriptions.
“This space has been ripe for a makeover for a long time,” Brendan Alberts, head of search and commerce at ad buying firm Dentsu, told The Wall Street Journal.
Nearly half of queries on Perplexity lead to follow-up questions, and the platform allows brands to sponsor answers. These responses can be “pre-approved and locked, giving you comfort about how your brand is represented in your responses.”
This is truly terrible.
Nearly 60% of U.S. consumers used chatbots to help make purchasing decisions in the past month. This is why platforms are trying so desperately to cash in on their results.
subscribe
The most interesting read on blockchain. Delivered once a week.
Andrew Fenton
Andrew Fenton, based in Melbourne, is a journalist and editor covering cryptocurrency and blockchain. He has worked as a national entertainment writer for News Corp Australia, a film journalist for SA Weekend and The Melbourne Weekly.
Follow the author @andrewfenton