Week 38

View this email in your browser Hi Humans 👋🏻 It’s a no-brainer that this newsletter will cover the launch of OpenAI’s new LLM model. However, OpenAI is not going to have it all, as we’re also going to explore how robo-taxis will change road safety forever 🚕🍓 But first, a quick dive into AI news of the last week: Slack -  is turning into an AI agent. Incogni will remove your online data. Is prompt engineering dead? Figmas AI update is being called a fiasco. OpenAI Strawberry OpenAI has once again made headlines with their latest project – the o1 model, also known as Strawberry 🍓 This cutting-edge AI isn't just about giving quick answers; it’s designed to think and reason before responding - like the human brain. The o1 model, the first in a series of “reasoning” models, promises to handle complex questions more effectively than before. From solving intricate math problems to writing code for new video games, o1 could change the game in numerous fields. Think of it like having a super-smart assistant that not only provides answers — but also shows you how it’s thinking through a problem, step-by-step. One of the more eyebrow-raising aspects of the new model is its capability to "scheme." Yes, you read that right. While the idea of a reasoning AI sounds fantastic, it brings along concerns about safety and trust. Why It Matters 🌍 The development of AI’s like o1 could have monumental effects on our future. It holds the potential to revolutionize fields like medicine and engineering with improved problem-solving abilities. But with great power comes great responsibility. AI experts like Dan Hendrycks and Yoshua Bengio have voiced their concerns, pointing out that advanced reasoning abilities could make AI more dangerous if not properly kept in check. This has led to what’s being called the "safety paradox": sometimes improving the safety of AI in one area means accepting that it might be less safe in other respects. Hey you - join us here 👉🏻 Be safe, choose the robo-taxi 🚕 Waymo, a leader in self-driving technology owned by Alphabet (Google), has revealed some fascinating insights: while Waymo’s robo-taxis are very safe, human drivers are responsible for the majority of all serious collisions.   Even in the bustling streets of San Francisco and Phoenix, these driverless cars have a much better safety record than human drivers. Key Numbers 22 million miles driven by Waymo's driverless vehicles. Less than one injury-causing crash per million miles for Waymo, compared to 64 crashes per million miles for typical human drivers. In serious crashes that trigger an airbag, Waymo self-driving cars are six times safer. Source: Waymo Why It Matters 🌍 While these numbers are encouraging, it's important to remember that self-driving technology isn't perfect yet. It currently performs best in controlled environments and will need even more sophisticated systems to manage the unpredictability of human drivers and complex traffic scenarios. However, the key takeaway is profound: the true promise of self-driving technology lies in its potential to drastically reduce accidents and save lives. Waymo's data shows we're on the right track, but careful development is crucial. As this technology evolves, we can anticipate a future where commutes are not just faster or more convenient but significantly safer. Join our community here: Thank you for reading! If you have inputs, or improvements - please reach out to me on LinkedIn:) This mail is made in collaboration with Martin Rosengaard and Silje Buch Larsen. Until next time // Annemette Møhl Copyright (C) 2024 AI.Hub CPH. All rights reserved. You are receiving this email because you opted in via our website. Our mailing address is: AI.Hub CPH Hojbro Plads 10 Kobenhavn K 1200 Denmark Want to change how you receive these emails? You can update your preferences or unsubscribe