Open in app
Sign inGet started
The Turing Test

The Turing Test

… its passing is greatly exaggerated
Go to the profile of John Ball
John Ball
AGI needs the scientific method

AGI needs the scientific method

… but today’s models currently block it
Go to the profile of John Ball
John Ball
Why LLMs aren’t AGI

Why LLMs aren’t AGI

Transformers fundamental compromise of accuracy
Go to the profile of John Ball
John Ball
Thinking about a Thought Experiment

Thinking about a Thought Experiment

What would John Searle or Alan Turing say?
Go to the profile of John Ball
John Ball
LLMs missing Language

LLMs missing Language

‘words’ are not just symbols, or vectors. It’s meaning in context.
Go to the profile of John Ball
John Ball
What is needed for AGI by 2029?

What is needed for AGI by 2029?

a SCIENTIFIC change, such as using lossless knowledge
Go to the profile of John Ball
John Ball
What’s wrong with JUST generating text?

What’s wrong with JUST generating text?

a killer app needs MEANING, too
Go to the profile of John Ball
John Ball
Analyze Brain Regions, not Neurons

Analyze Brain Regions, not Neurons

where today’s AI went wrong
Go to the profile of John Ball
John Ball
Introducing Deep Symbolics

Introducing Deep Symbolics

the next step from deep learning towards AGI
Go to the profile of John Ball
John Ball
How ‘2+2=4’ works for better AGI

How ‘2+2=4’ works for better AGI

Explaining pattern-based arithmetic in detail
Go to the profile of John Ball
John Ball
About Pat IncLatest StoriesArchiveAbout MediumTermsPrivacyTeams