AI EXPLOSION: GEMINI 1.5, GEMINI ADVANCED, OPEN AI SORA, & DISRUPTING INSURANCE
Dive into recent AI breakthroughs: Google Gemini 1.5's massive 1.5M token capacity and OpenAI Sora's realistic video generation. Explore how AI adoption differs between insurance incumbents and startups, and consider the future integration of AI in daily workflows, from Venture Step.
TL;DR
Gemini 1.5 handles 1.5M tokens. Sora turns text into stunning video. Insurance is next in line to be reshaped.
THE LEAP FORWARD IN AI CAPABILITIES
In the rapidly evolving landscape of artificial intelligence, we've witnessed remarkable advancements reshaping how we approach content creation, data analysis, and business operations. As the Venture Step Podcast host, I've closely followed these developments, particularly the recent releases of Google's Gemini 1.5 and OpenAI's Sora.
GEMINI 1.5 BREAKING TOKEN BARRIERS
One of the most significant breakthroughs comes from Google's Gemini 1.5, which can handle an impressive 1.5 million tokens. To put this in perspective, the average book contains around 300,000 tokens, and ChatGPT turbo allows 128,000 tokens. This massive increase in token capacity enables a deeper understanding and analysis of extensive content. Through early access testing, researchers have reported 99.7% accuracy on a million tokens, demonstrating the model's exceptional reliability. This capability was showcased through various impressive demonstrations, including:
- Analysis of hour-long NBA dunk contest videos, identifying and explaining the highest-scoring dunks
- Translation into rare languages like Saterland Frisian, spoken by fewer than 2,000 people
- Ability to process one-hour videos, 11 hours of audio, and over 30,000 lines of code
Shows the token increase for Gemini 1.5 Pro
SORA: THE NEXT FRONITER IN AI-GENERATED VIDEO
OpenAI's Sora represents a quantum leap in AI-generated video capabilities. The progress is striking - just 10 months ago, we looked at distorted, unsettling videos like the infamous "Will Smith eating spaghetti" clip. Today, Sora produces remarkably realistic content from simple text prompts, including:
- Photorealistic scenes of people in various settings, complete with accurate reflections and lighting
- 3D animated characters with Disney-like quality
- Historical recreations, like colorized Gold Rush footage
- Complex environmental simulations with detailed physics and lighting effects
DISRUPTING THE INSURANCE INDUSTRY: A TALE OF TWO APPROACHES
Research Paper - Artificial intelligence (AI) and Business innovation in Insurance: A comparison of incumbent firms versus new Entrants by Christopher P. Holland and Anil S. Kavuri
The impact of AI extends beyond content creation into traditional industries like insurance. Based on my experience in the sector, I've observed a fascinating divide between incumbents and new entrants in their AI adoption strategies:
Artificial intelligence (AI) and Business innovation in Insurance: A comparison of incumbent firms versus new Entrants by Christopher P. Holland and Anil S. Kavuri. Incumbent Companies:
- Focus on enhancing existing workflows and customer service.
- Deal with legacy systems and complex data migration challenges.
- Implement AI for fraud detection and personalized offerings.
- Take a cautious, systematic approach to integration.
New Entrants:
- Prioritize customer-first products and innovative services.
- Target underserved markets with AI-driven solutions
- Implement automated claims processes.
- Offer dynamic pricing models using Internet of Things data.
THE FUTURE OF WORK AND AI INTEGRATION
As these technologies evolve, we're moving toward a future where AI becomes integral to every tech stack and daily workflow. While complete automation isn't imminent, the efficiency gains are undeniable. Having worked with various policy systems and data structures in insurance, I've seen firsthand how AI can streamline complex processes that once took months to complete. The key isn't about replacing humans but augmenting their capabilities. An enterprise AI subscription might be more cost-effective than maintaining large teams for specific tasks, but the human element remains crucial for strategic decision-making and creative direction.
CALL TO ACTION
Want to explore these topics further? Listen to the full episode, available on all major podcast platforms. Share your thoughts and experiences with AI adoption in your industry—I'd love to hear your perspective on how these technologies reshape your work. If you want to view the episode in video format, watch it on YouTube!
Listen
WATCH ON YOUTUBE
Support
AI Rundown founder Rowan Cheung’s early
https://x.com/rowancheung/status/1759280384930459941?s=20
OpenAI Sora – Script to Screen Magic demos
I ran all of the Sora prompts through Midjourney
— Nick St. Pierre (@nickfloats) February 16, 2024
Interesting how similar some are
side-by-sides against vids:
Open AI Paper
https://openai.com/research/video-generation-models-as-world-simulators
Research Paper
Holland, Christopher, and Anil Kavuri. "Artificial intelligence (AI) and business innovation in insurance: A comparison of incumbent firms versus new entrants." (2023).
Christopher P. Holland Loughborough University c.p.holland@lboro.ac.uk
Anil S. Kavuri Loughborough University a.s.kavuri@lboro.ac.uk