bigger than taylor and travis

gm,

Did you hear? AGI has been achieved internally at Boys Club.

love,

The Boys

Writer: SamEditor: Deana

It’s giving multimodal. Taylor and Travis aren’t the only ones stirring the rumor mill. Google has started privately testing its next-generation model, Gemini. Word on the street is it’s even more powerful than competitor OpenAI’s ChatGPT, and is set to launch later this year. The game changer is that it’s multimodal, meaning it can analyze various modes of input at the same time such as text, images, and voice. Not to be upstaged, ChatGPT just announced it would begin to roll out its own multimodal capabilities, letting users do things like upload a picture of a broken object and get instructions for how to fix it. We give it 0.0003 seconds until the first dick pic is uploaded.

A little help from our friends. The United Nations is considering tapping AI to help boost its slow progress towards its 2030 Sustainable Development Goals. Progress in areas such as education, climate change, and food security has been pretty mid so far, and stands to see vast improvements with the help of AI (as Billy G called out earlier this year). The UN is planning on announcing an AI advisory body that will identify the opportunities and risks of leveraging AI. We imagine that this will be the tenor of those debates.

Yes, chef. Researchers at the Toyota Research Institute created a new method of quickly teaching robots new skills. The approach involves giving them a sense of touch through a special mechanical blob-thing, using a new AI model, and then giving a demo on how the task is completed that the robots can “feel.” The first application of this groundbreaking invention? Breakfast. These robots can scramble eggs, flip pancakes, and complete 60+ other culinary tasks. The researchers also shared their vision to get robots to learn new behaviors on their own, and then teach them to a fleet of other robots. Cool!

We hang out in AI-image bot chatrooms so you don’t have to.