- AI Odyssey
- Posts
- NYC ❤️ AI
NYC ❤️ AI
...New York wants to be AI's world capital...
Welcome to AI Creator! 👋
Move over filters and Photoshop, a growing trend among young adults on platforms like TikTok involves using AI-powered apps like Remini, Try It On AI, and AI Suit Up to effortlessly generate professional-looking profile photos for job-seeking purposes on LinkedIn, with users praising the low-cost alternative to traditional professional headshots, while critics argue about the potential impact on self-esteem and authenticity in recruitment processes.
⏰ Top 5 Headlines
![](https://media.beehiiv.com/cdn-cgi/image/fit=scale-down,format=auto,onerror=redirect,quality=80/uploads/asset/file/621b6a64-ab46-41b8-a29f-2f4f41fa3e85/SEI_1758393131.jpg)
🔗Link: AI deciphers unopened ancient Roman scroll charred by volcanic eruption 2,000 years ago.
🔗Link: YouTube gets new AI-powered ads that let brands target special cultural moments.
🔗Link: Top £32,000-a-year private school appoints an AI chatbot as its new 'headteacher' in UK first.
🔗Link: AI could cause a financial crash within a decade, SEC head says.
🔗Link: Google promises to take the legal heat in users’ AI copyright lawsuits.
💬 Deeper Dives
New York wants to be AI's world capital
![](https://media.beehiiv.com/cdn-cgi/image/fit=scale-down,format=auto,onerror=redirect,quality=80/uploads/asset/file/ba4069a2-677c-40fa-ab75-d9cca92e78e4/image.png)
In a bid to dethrone the Bay Area as the global epicenter of technology, New York-based tech firms and investors are positioning the city as the world capital of AI. As part of this ambition, New York is set to host a 370-event Tech Week, starting October 16, organized by VC firm Andreessen Horowitz. The focus is on winning adoption for generative AI in key industries such as finance, communications, media, law, and medicine, which are concentrated in the New York area. Despite the West Coast's established tech community, industry leaders argue that New York's unique blend of customer proximity and diverse opportunities makes it an increasingly attractive hub for transformative AI tech.
Millions of workers are training AI models for pennies
![](https://media.beehiiv.com/cdn-cgi/image/fit=scale-down,format=auto,onerror=redirect,quality=80/uploads/asset/file/001114e4-66ff-4e20-90b3-c55c26d012fa/OSKARINA_Col_17_CCISO.jpg)
In a global trend from the Philippines to Colombia, millions of workers are toiling for meager pay to label training data for AI models employed by tech giants such as Amazon, Facebook, Google, and Microsoft. Oskarina Vero Fuentes, a Venezuelan worker for Australian data services company Appen, shares her experience of earning between 2.2 cents to 50 cents per task, working over 18 hours a day to make ends meet. The hidden industry, valued at $2.22 billion in 2022, is projected to reach $17.1 billion by 2030, raising concerns of data colonialism and exploitation in developing nations. Workers, facing uncertainty and low compensation, are pushing for industry unionization and recognition as valuable contributors to technological advancement.
Meta's bold open-source move
Meta's decision to open-source its Llama 2 AI model in July, described by a former top Facebook engineer as a move fueled by Mark Zuckerberg's daring spirit, has propelled the model to massive popularity. Despite concerns about potential misuse by malicious actors, the (mostly) open-source approach allows users to freely use, inspect, update, and redistribute the AI model—a concept integral to the tech industry since the 1990s. Although Llama 2 falls short of complete openness, developers appreciate its flexibility compared to closed models from giants like OpenAI and Google. With 30 million downloads as of late September, Llama 2 has become Zuckerberg's next major platform, showcasing Meta's bold stance in an industry that, according to the former Facebook engineer, often lacks courage in embracing open-source AI models.
To watermark AI, it needs its own alphabet
Distinguishing between AI-generated and human-generated content has become increasingly difficult, raising concerns about the misuse of AI, including scams, deepfakes, and automated spam. Recent incidents highlight the urgent need for a universal solution.
Unicode as the answer to AI watermarking (Proposed solution): Use Unicode, the universal numbering system for text, to create a distinct character set for AI-generated content. Unlike traditional watermarking methods, Unicode provides a fine-grained and universally applicable approach that works seamlessly across devices and platforms, offering a practical step towards transparency in the AI era.
Benefits and limitations of Unicode hack: Unicode-based AI watermarking could revolutionize how we verify the origin of digital content, with potential applications in consumer awareness, software development, government procurement, and education. However, challenges such as limited space in Unicode and compatibility with diverse languages need to be addressed for widespread adoption and effectiveness.
🧰 Startups
🔗Hook: Hook wants to help you create a legal remix of your favorite track for TikTok.
🔗Creative Forces: Raises $8.9m in Series A funding to transform e-commerce content production with generative AI.
🤳 Extra reads
Burned-out parents seek help from a new ally: ChatGPT. (Axios)
OpenAI's growth this year has been unstoppable. Maintaining it will be the hard part. (Insider)
Why AI is forcing startups to chase a mirage. (Inc.)
How roboticists are thinking about generative AI. (TC)
I tried the viral decorating AI RoomGPT on my boring living room. It added a pool in my apartment. (Insider)
The team behind today’s issue: Lavena Xu-Johnson, Hubert Trinkunas and Rosa Cecilia.