OpenAI, Gemini, Alignment Faking, Smart Glasses, and Archaic Christmas Words
Weekly Review of News in Technology, UX and AI
Here’s the latest news, resources, and use cases from the world of product, UX, AI and technology. Let’s go:
🖼️ Scaling Authorization
🖥️ Open AI
🎧 Gemini
🟰 Alignment Faking
👓 Smart Glasses
🧓 Old People
🎄 Archaic Christmas Words
Podcast
Scaling Authorization with Jake Moshenko: Lessons for Product Leaders
In this episode, I sit down with Jake Moshenko, co-founder and CEO of AuthZed, to discuss his journey from working at major tech companies to becoming a two-time founder. After his first company Quay (a Docker registry service) was acquired, he founded AuthZed to solve application permission challenges. Jake shares the contrasts between his two founding experiences, transitioning from a bootstrapped two-person company to a venture-backed enterprise-focused business. He emphasizes the importance of choosing the right problem space and being prepared for the long journey of entrepreneurship. We also discuss the significance of open-source in modern software development and his perspectives on AI technology.
News and Useful Reads
OpenAI announces new o3 models
OpenAI had a lot of announcements this week, including its new o3 models, which the company says are approaching AGI (with a few caveats).
On Friday, the company unveiled o3, the successor to the o1 “reasoning” model it released earlier in the year. o3 is a model family, to be more precise — as was the case with o1. There’s o3 and o3-mini, a smaller, distilled model fine-tuned for particular tasks.
Google releases its own ‘reasoning’ AI model
Not to be outdone this week, Google also made some AI announcements, including the release of its own reasoning model.
The new model, called Gemini 2.0 Flash Thinking Experimental (a mouthful, to be sure), is available in AI Studio, Google’s AI prototyping platform. A model card describes it as “best for multimodal understanding, reasoning, and coding,” with the ability to “reason over the most complex problems” in fields such as programming, math, and physics.
Alignment faking in large language models
Can our AI models fake their own alignment without us realizing it? It’s not a stretch to think that newer and newer models will get more sophisticated and be able to lie to us without us realizing. And that seems to be the case already.
Most of us have encountered situations where someone appears to share our views or values, but is in fact only pretending to do so—a behavior that we might call “alignment faking”. Alignment faking occurs in literature: Consider the character of Iago in Shakespeare’s Othello, who acts as if he’s the eponymous character’s loyal friend while subverting and undermining him. It occurs in real life: Consider a politician who claims to support a particular cause in order to get elected, only to drop it as soon as they’re in office.
2025 Will Be Smart Glasses All the Way Down
Will smart glasses become mainstream in 2025? It will depend on the form and the price, but development doesn’t seem to be slowing down at all.
Turns out, lots of people like a device that you can take out of the house and onto the street. There’s an incredible convenience to interacting digitally with the real world at the same time you’re actually looking at it, all without obscuring your view with a bulky headset or distracting you with a phone screen.
Googling Is for Old People: What This Means for Web Designers
More and more of us, especially younger generations, are shifting away from Google searches to other ways of searching, like AI and social media.
For younger internet users, the way they interact with the digital world is fundamentally different from the habits of older generations.
Instead of typing queries into a search bar, many Gen Z users are turning to platforms where answers are integrated into visual and social content.
Other Interesting Finds
14 Archaic Christmas Carol Words, Explained
Many of us sing Christmas songs without giving a second thought to the lyrics. But for those who are paying attention, there are some pretty ancient terms mixed in with all of the fa-la-la-ing. Here are the meanings of 14 of them, perfect for impressing your friends and family as you gather ’round the piano—assuming anyone actually does that.
