Another AI sorcery in the making: Gemini by Google DeepMind to rival ChatGPT
In April, Google merged its AI lab DeepMind with Google Brain (its AI research division) to form what’s today known as Google DeepMind. Eight weeks later, the new formation is making an announcement. In a piece by Wired, Demis Hassabis, DeepMind’s co-founder and CEO, drops the mic announcing that his team is working on a new AI system.
The AI system will be named Gemini and it aims to beat ChatGPT in the AI competition. How? By using techniques from AlphaGo (a DeepMind software, that made 2016 headlines after beating a champion in the complex board game Go), says Hassabis:
AlphaGo utilizes a ‘reinforcement learning’ technique developed and pioneered by DeepMind: the software learns from feedback that it gets from performing repetitive tasks. For example, when in the start-up stage, DeepMind was trained in retro games like Breakout, Pong and Space Invaders. Starting with zero knowledge about the given game, the software is in repetitive mode: it makes moves, makes decisions and learns about whether it’s getting ‘ahead’ in the game or not.
While humanity as a whole still struggles to exhaust the list of all the things one could do with AI, one particular concern rose in the last year: ChatGPT can suck at math problems, reddit users report. Google DeepMind CEO and his team are ‘aiming to give the system new capabilities such as planning or the ability to solve problems’, which will be an interesting turn of events, if executed correctly. The AI division is also looking for ways to enhance the LLM technology (large language model - the deep learning algorithm that can recognize, summarize and generate text) with ideas from other areas of AI.
At the end of May, a statement by the Center for AI Safety was signed by many notable names in the AI field, like OpenAI’s Sam Altman, DeepMind CEO Demis Hassabis, MIT’s Max Tegmark, Microsoft CTO Kevin Scott, etc. ‘Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war’, reads the statement.
Let’s wrap this up with a catchphrase from ‘The Hitchhiker's Guide to the Galaxy’, the one that is written in ‘large, friendly letters on the cover’: ‘Don’t Panic’. Long before (if at all) we get to the ‘Man vs. Machine’ match, we’ll witness the ‘Gemini vs ChatGPT’ spectacle.
At a high level, you can think of Gemini as combining some of the strengths of AlphaGo-type systems with the amazing language capabilities of the large models. We also have some new innovations that are going to be pretty interesting.
The project is still in development and will be for many months to come. It could cost up to hundreds of millions of dollars, Hassabis also says. For reference, OpenAI’s CEO (Sam Altman) recently disclosed that over $100 million were spent to create GPT-4.
Why AlphaGo?
AlphaGo utilizes a ‘reinforcement learning’ technique developed and pioneered by DeepMind: the software learns from feedback that it gets from performing repetitive tasks. For example, when in the start-up stage, DeepMind was trained in retro games like Breakout, Pong and Space Invaders. Starting with zero knowledge about the given game, the software is in repetitive mode: it makes moves, makes decisions and learns about whether it’s getting ‘ahead’ in the game or not.
What else is up Gemini’s sleeve?
While humanity as a whole still struggles to exhaust the list of all the things one could do with AI, one particular concern rose in the last year: ChatGPT can suck at math problems, reddit users report. Google DeepMind CEO and his team are ‘aiming to give the system new capabilities such as planning or the ability to solve problems’, which will be an interesting turn of events, if executed correctly. The AI division is also looking for ways to enhance the LLM technology (large language model - the deep learning algorithm that can recognize, summarize and generate text) with ideas from other areas of AI.
AI could cuddle somewhere between ‘pandemic’ and ‘nuclear war’
At the end of May, a statement by the Center for AI Safety was signed by many notable names in the AI field, like OpenAI’s Sam Altman, DeepMind CEO Demis Hassabis, MIT’s Max Tegmark, Microsoft CTO Kevin Scott, etc. ‘Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war’, reads the statement.
Let’s wrap this up with a catchphrase from ‘The Hitchhiker's Guide to the Galaxy’, the one that is written in ‘large, friendly letters on the cover’: ‘Don’t Panic’. Long before (if at all) we get to the ‘Man vs. Machine’ match, we’ll witness the ‘Gemini vs ChatGPT’ spectacle.
FCC OKs Cingular\'s purchase of AT&T Wireless
Things that are NOT allowed: