A week in Generative AI: Copilot, Elections & Geometry
A week in Generative AI: Copilot, Elections & Geometry
News for the week ending 21st January 2024
If there’s one thing to take away from this week’s newsletter, it’s how profound the impact of Microsoft’s Copilot will be for people using Microsoft365 in their day-to-day jobs. Watch the video below and you’ll see what I mean!
Consumers can now use Microsoft’s Copilot in Word, Excel, Powerpoint etc.
This is the first of what I call big, enterprise applications of generative AI technology. The demo of PowerPoint above by Ethan Mollick is very impressive - a whole powerpoint built from just a simple 8 word prompt.
How OpenAI is approaching 2024 worldwide elections
It’s good to see OpenAI make these commitments, but sadly I don’t think it will make a difference. Open Source models are freely available and the challenges of misinformation in politics can only be solved where the misinformation is distributed i.e. the social platforms.
DeepMind’s latest AI can solve geometry problems
Another small step towards generative AI models being able to reason, but I see this as more of a ‘brute force’ approach to intelligence rather than anything more elegant. However, this reinforces my view that we have all the ingredients for AGI, we just haven’t worked out how to put them all together yet.
‘Very scary’: Mark Zuckerberg’s pledge to build advanced AI alarms experts
This seems to be just a continuation of Zuckerberg’s quest to create a more connected and transparent world, regardless of the consequences. However, the stakes are so much higher with AGI, and I don’t think these commitments should be thrown around lightly when we don’t understand the consequences, either intended or unintended.
Altman says ChatGPT will have to evolve in “uncomfortable” ways
This is the first time Altman has talked about the personalisation of generative AI and it’s great to hear as it is intuitively a big factor in the future of the technology manifesting as an ‘assistant’. There have been peaks of this in some recent leaks - Excited for what’s coming!
AI Ethics News
OpenAI, Microsoft’s Partnership May Be Killed Off Before the Dawn of AGI
New study from Anthropic exposes deceptive ‘sleeper agents’ lurking in AI’s core
Google AI has better bedside manner than human doctors — and makes better diagnoses
Long Reads
One Useful Thing - The Lazy Tyranny of the Wait Calculation
Unconfuse Me - Sam Altman & Bill Gates discussion
“The future is already here, it’s just not evenly distributed.“
William Gibson