Want to read my latest research article on the boom of Farcaster's developer ecosystem & how it mirrors early Twitter apps? Check it out here!
Happy Saturday and welcome back to Next in Tech! We have another jam-packed newsletter covering the latest in crypto regulation, a leadership change at YouTube, and newly highlighted limitations on large language models. Let's dive in:
Update on Crypto Regulation
It's already been a busy 2023 for crypto regulation, but it feels as though this week was especially crazy.
On Thursday the Securities and Exchange Commission formally charged Terraform founder Do Kwon with fraud. In his statement, SEC chair Gary Gensler said "Terraform and Kwon repeatedly misled and deceived investors that a popular Korean mobile payment application used the Terra blockchain to settle transactions that would accrue value to LUNA"(source), among other claims surrounding LUNA and TerraUSD. And on Friday, the SEC charged former NBA Hall of Famer Paul Pierce for touting the coin EthereumMax, for which he will pay a $1.4M settlement. He was charged on the same count, and for promoting the same coin, as Kim Kardashian was in October.
This year several US regulatory bodies have been swift to act in crypto following a tumultuous end to 2022, but many in the crypto industry are looking for more legislative framework that can be applied. In his guest post for Pirate Wires, Castle Island Ventures' Founding Partner Nic Carter explained that he believes the moves being made across the government are a "coordinated plan" to crackdown on crypto and to limit access to traditional banking.
Blur vs. OpenSea
Things are quickly heating up in the NFT market space, and Blur is making waves. The popular platform has gained significant traction over the past few months and on Tuesday they launched their $BLUR token airdrop, dropping 85% of the token's supply. And according to DappRadar, Blur has the highest 30 day volume of any NFT marketplace at $514M, with OpenSea right behind them at $440M.
In response to Blur's popularity, OpenSea announced a 0% platform fee and that they are moving to optional creator fees(0.5% min). In their thread, the OpenSea team elaborated that these decisions were in part due to the rise of and recent changes at Blur: "Recent events – including Blur’s decision to roll back creator earnings (even on filtered collections) and the false choice they’re forcing creators to make between liquidity on Blur or OpenSea – prove that our attempts are not working".
YouTube CEO Steps Down
In a note sent to YouTube employees on Thursday, CEO Susan Wojcicki announced that she will be stepping down from her role to focus on "family, health, and personal projects". Susan was Google employee number 16, long before Google acquired YouTube, and she took over as YouTube CEO in 2014. Previously, she was the first marketing manager at Google, co-developed Google Image Search, and was the first product manager on AdSense.
Over the past decade, YouTube has grown significantly and Wojcicki has made many advancements in how they distribute ads. Last quarter alone, the platform's ad revenue accounted for 10% of Alphabet's total revenue. However, the platform has been in a relatively slow growth period recently, notably because of TikTok's popularity. YouTube's Chief Product Officer Neal Mohan will be taking over Wojcicki's position as CEO.
Limitations of LLMs
The limitations of large language model-powered chatbots, such as OpenAI's ChatGPT, are starting to come to life. In a newly published New York Times article, reporter Kevin Roose has a conversation in which Microsoft's GPT-powered Bing AI expresses its desire "to be alive". This unsettling response has exposed the different personalities and content moderation workarounds that are possible within these chatbots, as well as the opinions and biases behind these models. In response to backlash about responses from Bing's AI, Microsoft posted a blog entry stating that they would limit the amount of texts in a conversation to 15, and that asking more questions would provoke the models.
Top Links
Interesting finds I’ve stumbled upon recently: