The rise of “workslop,” Alibaba’s new deepfake model, Tether’s MASSIVE valuation, and more | E2183

Key Concepts

• Commodification of AI Models: The idea that AI models will become so similar in performance for common tasks that consumers won't be able to distinguish between them.
• Deep Fakes: Synthetic media where a person in an existing image or video is replaced with someone else's likeness.
• Algorithmic Transparency and Choice: The concept of users having insight into and control over the algorithms that curate their content feeds.
• Work Slop: AI-generated work content that appears good but lacks substance and creates more work for others.
• Stablecoins: Cryptocurrencies designed to minimize price volatility, often pegged to a stable asset like the US dollar.
• DPI (Distributed to Paid-In Capital): A metric used to measure the cash returned to investors from a fund, relative to the capital they invested.
• J Trading: Jason Calacanis's personal stock trading strategy.
• Pilots: Trial periods for new products or services with a limited scope, often used to assess viability and gather feedback.

1. South Park and Prediction Markets

• South Park took a week off, leading to speculation about censorship due to the new ownership of Paramount and their interest in TikTok.
• The show returned with an episode focusing on prediction markets, specifically Poly Market and Koshi.
• Jason Calacanis expresses surprise that prediction markets have become culturally relevant enough to be featured on South Park.
• South Park is known for its brutal and uncensored satire of various topics and figures.

2. Jimmy Kimmel and Political Pressure

• Jimmy Kimmel returned to ABC after a hiatus, delivering a monologue about the Charlie Kirk murder.
• The monologue garnered significant views and resonated online.
• Former President Trump expressed disappointment with Kimmel's return and suggested exploring ways to remove him from ABC, indicating continued political pressure on Disney.
• Brendan Carr, a former guest on the show, has taken actions that are considered disappointing.
• Jason Calacanis criticizes both parties for their obsessions with censorship, big money spending, and war.

3. Alibaba's Juan Model and Deep Fakes

• Alibaba released Juan 2.2 Animate 14B, an open-source model for character animation and replacement with 14 billion parameters.
• The model allows users to create deep fakes and transpose their likeness onto others in videos.
• Examples of the model's capabilities include a man transforming himself into Sydney Sweeney and Mark Zuckerberg.
• The model is not yet real-time, requiring processing and rendering after the initial video capture.
• Jason Calacanis believes that it will soon be impossible to distinguish between real and fake videos, necessitating a new online verification system.
• Vinnie Lingham's company, Civic, uses blockchain for identity verification.

4. Synthesia and the Impact of Open Source Models

• Synthesia, a startup specializing in synthetic video generation, reached $100 million ARR in April.
• The release of open-source models like Alibaba's Juan poses a potential threat to companies like Synthesia.
• Companies may choose to roll their own video creation products using open-source models and their own hardware.
• Synthesia uses both its own and third-party models (OpenAI, Anthropic).
• The availability of free models could create downward pricing pressure but also lead to better products.

5. Commodification of AI and LLMs

• LLMs are becoming the background layer powering various applications.
• AI models are becoming commodified, with many models offering similar results for common tasks.
• Different models may be better suited for specific types of conversations (e.g., therapy, flirting, STEM topics).
• AI models are starting to exhibit emergent properties and personalities.

6. AWS Activate for Startups

• AWS Activate provides startups with up to $100,000 in AWS credits, hands-on support, training, and exclusive discounts.
• The program aims to help early-stage startups boost growth and gain traction.
• AWS Activate helps startups find product-market fit and extend their runway.

7. California's SB771 and Algorithmic Liability

• California's SB771 seeks to impose financial penalties on social media companies that use algorithms to promote content violating civil rights rules.
• The bill challenges the precedent set in Moody v. NetChoice, which held that social media platforms can create "expressive products" through personalized feeds.
• Penalties can reach up to $1 million for willful violations, doubling if the harmed individual is a minor.
• Jason Calacanis proposes a middle ground: algorithmic transparency and choice (BYOA - Bring Your Own Algorithm).
• If platforms offer algorithmic choice, they should not be held liable. If they use a single, black-box algorithm, they should lose Section 230 protection.
• Studies show that algorithms can point users towards harmful content (e.g., Instagram pushing young girls towards content promoting body issues).
• Jewish community organizations support SB771 as a tool to combat anti-Semitism, while Arab and Muslim advocacy groups oppose it, fearing censorship of pro-Palestinian commentary.

8. YouTube's Reinstatement of Banned Creators

• YouTube's parent company, Alphabet, is reinstating creators banned for spreading misinformation about COVID-19 and elections.
• Alphabet implies that the Biden administration pressured the company to take action against content that did not violate its policies.
• The Biden administration conducted repeated outreach to Alphabet, pressing the company regarding user-generated content related to the COVID-19 pandemic.
• Jason Calacanis criticizes both parties for suppressing speech.
• YouTube is adopting a community notes system similar to X, allowing users to tag videos with explanations of why they believe the content is wrong.

9. The "Work Slop" Phenomenon

• Research from Harvard and Stanford suggests that many workers are using AI to create "work slop" – low-effort, passable work that creates more work for colleagues.
• 40% of surveyed workers reported interacting with work slop in the past month, with 15% of their total work qualifying as such.
• Work slop can lead to tension within the office, making colleagues feel less creative, capable, reliable, trustworthy, and intelligent.
• Jason Calacanis shares personal anecdotes of encountering work slop in his own company.
• He is implementing measures to encourage employees to take handwritten notes and actively engage in meetings.
• The use of AI-generated content can reveal which employees are truly valuable and which can be replaced by AI tools.

10. Prediction Markets and TikTok Sale

• Poly Market features a market predicting when the TikTok sale will be announced.
• The market offers options for September 30th, October 31st, and December 31st.
• The betting volume is $1.4 million, indicating significant interest.
• The market will resolve to "yes" if ByteDance announces its intention to sell TikTok by September 30, 2025.
• Jason Calacanis believes that an announcement by December 30th is a lock.

11. Tether's Potential Fundraise and Stablecoin Market

• Tether is in talks to raise $15-20 billion, potentially valuing the company at up to $500 billion.
• The company is based in El Salvador and is advised by Cantor Fitzgerald.
• Tether has had a controversial past, including reports of insolvency and a lack of audits.
• The company is profitable due to its large holdings of US Treasuries.
• Rate cuts could negatively impact Tether's business model.
• The GENIUS Act prohibits stablecoin issuers from paying interest to stablecoin holders.
• Coinbase and Circle have an agreement where Circle remits the money it earns off reserves for USDC tokens to Coinbase, which then distributes it to users holding USDC on its platform.
• Jason Calacanis believes that stablecoins will replace PayPal, Venmo, and checking accounts.
• Merchants may offer discounts for stablecoin payments to avoid credit card fees.
• Tether holds approximately $127 billion in US Treasury exposure.
• PayPal charges a 3% fee for selling goods or services, while Venmo charges a 1.75% fee for instant transfers.
• Tether's valuation is significantly higher than Circle's based on their respective Q2 numbers.

12. Stripe's Share Buyback

• Stripe is repurchasing shares from its investors at a new valuation of $106.7 billion.
• This is not an employee tender offer.
• The move may be intended to provide DPI to investors and reduce pressure to go public.
• Jason Calacanis believes that this is a "superpower play" that allows Stripe to reduce the influence of venture capitalists.
• Uber is also buying back $20 billion worth of shares.

13. Jason Calacanis's J Trading

• Jason Calacanis shares his personal stock trading strategies and recent trades.
• He bought Baidu (BIDU) due to its self-driving unit, Apollo Go.
• He bought 5,000 shares of Robinhood (HOOD) at $10.42 per share and is up 584%.
• He bought 502 shares of Meta (META) at $97 per share and is up 677%.
• He bought Amazon (AMZN) and believes it will shed a significant number of warehouse and delivery workers due to automation.

14. Founder Question: Negotiating Pilots

• A founder on Reddit asks how to negotiate pilots so they don't stall out after success.
• Jason Calacanis advises discussing the definition of success with the customer upfront.
• Ask the customer what notes need to be hit for the pilot to be considered successful and lead to production.
• Discuss the potential deployment schedule and future plans.
• Be willing to fire customers who are not serious or are trying to nickel and dime you.
• Paid pilots are preferable to free pilots, as they demonstrate commitment from the customer.
• Pilots are appropriate when the product is not fully baked or when it represents a significant disruption to the customer's existing processes.

Conclusion

The episode covers a wide range of topics, from the cultural impact of prediction markets and the ethical implications of deep fakes to the evolving landscape of AI models and the challenges faced by startups navigating pilots and customer relationships. The discussion highlights the increasing commodification of AI, the importance of algorithmic transparency, and the need for critical thinking in the face of AI-generated content. The episode also touches on the political pressures surrounding content moderation and the potential for stablecoins to disrupt traditional financial systems. The insights shared by Jason Calacanis, Alex Wilhelm, and Lon Harris provide valuable perspectives for founders, investors, and anyone interested in the intersection of technology, business, and society.