University Associate Professor of Computer Science at Cambridge University, specialising in AI and Machine Learning. Player of many instruments.
Joined December 2021
- Tweets557
- Following137
- Followers138
- Likes220
AI Top Tip #15
It doesn't *do* anything: it has no intention, no motivation beyond what *people* are willing to allow it to do.
If AI does something bad, it's the *owner's* fault.
Right now, the main AI companies are a little like drunk drivers...
Three things I wish the BBC would stop doing:
1. Using strident influencers who know nothing about AI to comment on AI.
2. Anthropomorphising AI. AI will not hurt us, because it has no intent or motivation. Humans using it though...
3. Stop being in thrall to tech-bros.
After 30+ years as an AI researcher I still maintain the following thoughts:
1. What I don't want AI for: stuff I enjoy, like coding, music and mathematics.
2. What I do want AI for: cleaning the toilet and taking out the bins.
ukri.org/news/new-72-million…
Meanwhile, Jodrell Bank dies for want of £2.8m.
Sean Holden retweeted
Reminder: this is all fake. Any time someone makes a claim like this, either they are lying, or every single programmer they know is completely awful and can’t program at all.
teapawt.org/alexalbert__/status/19…
Since Claude 4 launch: SWE friend told me he cleared his backlog for the first time ever, another friend shipped a month's worth of side project work in the past 5 days, and my DMs are full of similar stories.
I think it's undebatable that devs are moving at a different speed now. You can almost feel it in the air that this pace is becoming the default norm.
When I was a child, a school-book contained the following poem:
Beware the fly
He rings his hands
He rings his feet
For some reason it has stayed with me.
I can't find any information about the poem or its author.
Can anyone help?
"The Algorithm" has decided to show me a lot of poetry.
The poetry, seen among the stream of utter bollocks, actually does restore my faith in humanity.
Please, please...
... more poetry.
Just to be clear, Altman didn't invent this technology any more than Musk invented electric cars.
The credit goes to John Daugman.
time.com/7288387/sam-altman-…
"Privacy preserving." ????
AI really has become nothing but an unwanted experiment in surveillance capitalism forced on the world.
Bring one of these near me, and watch me reprogram it with a very large hammer.
time.com/7288387/sam-altman-…
AI Top Tip #14
You don't know what "consciousness" is, or anything about the mechanism that generates it.
And neither does anybody else.
So if your definition of "AGI" requires "consciousness", stop spouting total guff about how it's nearly here.
Well, I never saw that co... sorry, can't keep a straight face... Phhhhwahahaha! Same build quality as a Tesla eh?
If you want to 'make life multiplanetary' you'll be needing one that doesn't keep exploding.
telegraph.co.uk/business/202…
AI Top Tip #13
AI researchers and AI companies are different.
AI researchers (the honest ones) want to communicate interesting, verified knowledge.
AI companies want to raise investor cash.
Remember this. Be particularly sceptical of advertisers pretending to be tech firms.
AI Top Tip #12
Don't give any information to a free LLM-based service, unless you'd be happy to tell it to your worst enemy - someone actively interested in f**king you up.
There's a reason these things are free...
Lots of people here don't realise that:
- You don't have to give your information to someone's LLM.
- If an LLM tells you to do something stupid, and you do, that's your problem.
- If an LLM provides bullshit, and you rely on it, that's your problem.
Try a little scepticism!
I love working at Trinity College Cambridge. For the last few hours a colleague with a Nobel Prize has been trying to get me to listen to Country Music. This will never happen. I've been trying to convert them to Jazz. We disagree and everyone's happy.
Sean Holden retweeted
🚨 New study reveals that when used to summarize scientific research, generative AI is nearly five times LESS accurate than humans.
Many haven't realized, but Gen AI's accuracy problem is worse than initially thought.
