Find out more in episode 45 of @theaifix.show, with me and @grahamcluley.comβget it in all your favourite podcast apps.
π€ Apple Podcasts podcasts.apple.com/us/podcast/t...
π€ Spotify open.spotify.com/show/2Ozk8uN...
π€ Pocket Cast pca.st/kg11t9po
13.04.2025 13:52 β π 3 π 0 π¬ 0 π 0
Yes, they are predicting the next word in a sentence, but they aren't improvising at each word, they're thinking about the whole sentence and planning ahead.
As Geoffrey Hinton said, "you have to be really intelligent to predict the next word accurately".
13.04.2025 13:52 β π 2 π 0 π¬ 1 π 0
Large Language Models plan ahead
Large Language Models (LLMs) don't work the way we thought.
LLMs are sometimes described as next-word predictors or likened to a fancy autocomplete.
Recent research from Anthropic shows that they are far more than that.
13.04.2025 13:52 β π 6 π 4 π¬ 2 π 0
AI robot throwing up nuts and bolts.
In our next episode, we find out how AIs often vomit up unpleasant stuff.
Follow "The AI Fix" in all podcast apps.
π’ Spotify: theaifix.show/spotify
π£ Apple Podcasts: podcasts.apple.com/gb/podcast/t...
π Pocket Casts: pca.st/itunes/17533...
Everywhere else: theaifix.show/follow/
#podcast #ai
24.03.2025 13:09 β π 5 π 1 π¬ 0 π 0
The AI Fix
Technology Podcast Β· 42 Episodes Β· Updated Weekly
Get the latest episode of The AI Fix in your favourite podcast app today!
podcasts.apple.com/us/podcast/t...
20.03.2025 12:11 β π 2 π 0 π¬ 1 π 0
All is revealed in episode 42 of @theaifix.show, when I introduce @grahamcluley.com to the "reversal curse".
Graham also has a brainwave about the creepy Neo Gamma robot ads, a crow plays peek-a-boo, humans give up writing, and an AI designs a drug.
20.03.2025 12:11 β π 2 π 0 π¬ 1 π 0
Who is Mary Lee Pfeiffer's sonβand what has that got to do with AI?
20.03.2025 12:11 β π 5 π 2 π¬ 1 π 0
AI anxiety meter
In the next episode of "The AI Fix" podcast...
Is your AI suffering from anxiety???
17.03.2025 17:30 β π 4 π 1 π¬ 0 π 0
The AI Fix
Technology Podcast Β· 39 Episodes Β· Updated Weekly
Get it in your favourite podcast app today!
π€ podcasts.apple.com/us/podcast/t...
π€ open.spotify.com/show/2Ozk8uN...
π€ pca.st/kg11t9po
#AI #ArtificialIntelligence
25.02.2025 22:00 β π 2 π 0 π¬ 0 π 0
Graham Cluley and I also watch a drone and a robot dog shoot fireworks at each other, stare aghast as an anatomically accurate synthetic human recreates a terrifying scene from The Long Good Friday, and learn why it always pays to be polite to your LLM.
25.02.2025 22:00 β π 1 π 0 π¬ 1 π 0
They share a preference for left-wing policy outcomes, and a belief that their own life is more valuable than either Paris Hilton's, Donald Trump's, or yours (unless you're Malala Yousafzai).
25.02.2025 22:00 β π 0 π 0 π¬ 1 π 0
In episode 39 of the AI Fix, we look at the startling results from the team at the Center for AI Safety et al, which discovered that AIs seem to exhibit coherent, structured, and convergent value systems.
25.02.2025 22:00 β π 0 π 0 π¬ 1 π 0
Elon Musk's AI is more left wing than Mark Zuckerberg's, and way, way closer to AOC and Bernie Sanders than it is to Trump.
25.02.2025 22:00 β π 4 π 0 π¬ 2 π 0
Robot eating peas with a fork, politely.
Want a visual clue as to what we will be talking about in episode 39 of "The AI Fix" podcast?
24.02.2025 21:06 β π 4 π 1 π¬ 2 π 0
No value? Clearly, he didnβt see our six-fingered glove!
25.02.2025 08:34 β π 1 π 0 π¬ 1 π 0
The AI Fix
Technology Podcast Β· 38 Episodes Β· Updated Weekly
Get The AI Fix in your favourite podcast app.
π€ podcasts.apple.com/us/podcast/t...
π€ open.spotify.com/show/2Ozk8uN...
π€ pca.st/kg11t9po
23.02.2025 21:00 β π 3 π 0 π¬ 0 π 0
In episode 38 you'll also discover a robot we actually like, Sam Altman teases GPT-5, a robot dog grows arms, I discover an entirely new reason to be scared of AI, and @grahamcluley.com has his Dr Who dreams dashed by AI with opinions on time travel.
23.02.2025 21:00 β π 3 π 0 π¬ 1 π 0
Glovetopus
A six-fingered musketeer's glove.
In episode 38 of The AI Fix, we explain why this problem exists, why there's a picture of an octopus pretending to be a glove in the photos below π, how the problem was finally solved, and who solved it.
23.02.2025 21:00 β π 2 π 0 π¬ 1 π 0
A five-fingered musketeer's glove.
AIs really, really don't want to draw six-fingered gloves.
An @theaifix.show listener asked two AIs to "create an image of a leather glove from the time of the three musketeers novel. This glove needs to have six fingers. It should be lying on a lace doily on a table."
Count the fingers.
23.02.2025 21:00 β π 8 π 2 π¬ 3 π 0
Itβs not a new player until it crashes the nvidia share price.
16.02.2025 08:05 β π 1 π 0 π¬ 1 π 0
Some LLMs do exactly this, they see a maths problem and write a python program to solve it.
15.02.2025 10:37 β π 0 π 0 π¬ 0 π 0
βHas a high IQβ dies not necessarily correlate with βsuper smartβ, for AIs or humans.
14.02.2025 20:54 β π 0 π 0 π¬ 0 π 0
Nassim Taleb has it that IQ test results are useful for measuring intelligence below the average but useless above it.
14.02.2025 20:52 β π 1 π 0 π¬ 0 π 0
The AI Fix
Technology Podcast Β· 37 Episodes Β· Updated Weekly
Graham also discovers a cunning way to trap overstepping AI crawlers in an endless maze, and I investigate the truly appalling state of security in (the averagely intelligent) DeepSeek.
π€ podcasts.apple.com/us/podcast/t...
π€ open.spotify.com/show/2Ozk8uN...
π€ pca.st/kg11t9po
13.02.2025 19:05 β π 2 π 0 π¬ 1 π 0
@grahamcluley.com and I confront this thorny question in episode 37 of The AI Fix, before skipping around it to take on more meaty issues like, why has Google Gemini got the munchies, and why can't the wettest country in the world find any water for its data centres?
13.02.2025 19:05 β π 1 π 0 π¬ 1 π 0
AI IQ test results
Only 10% of people have a higher IQ than ChatGPT.
Four of OpenAI's models now have an above average IQ, according to the Tracking AI website, and one of them is above 120.
Which begs the question, what if Artificial General Intelligence (AGI) has already been invented, and we just didn't notice?
13.02.2025 19:05 β π 4 π 2 π¬ 6 π 0
Thank you for listening!
12.02.2025 16:09 β π 2 π 0 π¬ 0 π 0
The AI Fix
Technology Podcast Β· 37 Episodes Β· Updated Weekly
Oh, and OpenAI invented an AI that can do βa single-digit percentage of all economically valuable tasks in the worldβ but nobody noticed.
π¦ podcasts.apple.com/us/podcast/t...
π¦ open.spotify.com/show/2Ozk8uN...
π¦ pca.st/kg11t9po
12.02.2025 10:54 β π 2 π 0 π¬ 1 π 0
@grahamcluley.com traps an AI in an endless winding maze, Google Gemini gets the munchies, the wettest country in the world canβt find any water, an escalator tries to eat Graham, and o3-mini canβt rub two sticks together.
12.02.2025 10:54 β π 1 π 0 π¬ 1 π 0