John Ball
@jbthinking
Followers
539
Following
2K
Media
65
Statuses
1K
I'm an inventor: scientist and engineer. My passion is in modeling the brains of animals to make machines more useful - with language. We're getting there!
Palo Alto, CA
Joined June 2010
Is the global expansion of electricity usage to support generative AI's sponsors a good idea? Many countries are budgeting billions to enable the new technology. It used to be that energy expansions were unsustainable.
1
0
0
RT @speechgenie_co: The brain starts learning from the moment you're born. Do you want to experience that hyper growth you did as an infant….
0
2
0
Learning is something a brain does, like this upwards spiral that keeps building on the lower levels!.
0
0
2
The weakness of the statistical linguistic models are their lack of sustainability. Sustainable AI runs on laptops and phones, not at giant datacenters fed with nuclear and unsustainable fuels as described in my book covering next generation techniques with Patom theory.
0
0
1
If we were nearing AGI at meta, wouldn’t the cost be irrelevant? If not nearing AGI what is the point of generative AI?.
Breaking 🫧 News: Meta freezes AI hiring and becomes the first major company succumb to the Great AI Bubble Deflation of 2025.
1
0
2
My substack article today looks at the common claim that brains are computational. That is a key reason AI is not progressing, especially in the hard areas like senses and motion for robotics. While I refer to Steven Pinker who was my go-to expert for.
0
2
6
Interesting update on ChatGPT, and the conclusion should be to give cognitive science a chance. Neurosymbolic models can be further researched and also brain models like Patom theory that already have implemented real time mapping from language to meaning and back using Role.
My work here is truly done. Nobody with intellectual integrity can still believe that pure scaling will get us to AGI. GPT-5 may be a moderate quantitative improvement (and it may be cheaper) but it still fails in all the same qualitative ways as its predecessors, on chess, on.
2
0
9
Which raises the question what is next to improve AI? The focus on computation and neural networks has had a long run since the 1940s but a better model will be based on cognitive science. And brain theory needs to align with what our brain does, which is unlike a computer!.
For years I warned that OpenAI might take a serious hit if GPT-5 was disappointing. We finally soon see whether I was right. In a rational world, their valuation would take a hit. • They no longer have anything like a clear technical lead. • GPT-5 is unlikely to be ahead of.
0
1
4
Is the brain computational? Does it use the information processing metaphor. It is difficult for many to find a brain function that isn’t like a computer, but it is needed for AI since brains hold the key to AI since they already work!.Psychology helps.
1
0
4
There are many reasons to look for alternatives to today’s big tech AI, but mainly so we can create AI systems that emulate humans. Language alone is a key need that is unlocked with the help from Role and Reference Grammar’s mature model.
I completely agree. And this is why it should be all hands on deck to find alternatives to LLMs.
0
2
2
Thanks Mary @CommBank. The bank claims that this fraud isn't covered by @Mastercard. The company was called Amazon KDP Publishers and we thought it was Amazon due to the branding. Attached is from the transcript where they claimed to be the internal department. We were tricked
@jbthinking Hi John, I am sorry to hear that this has happened to you. So we can further investigate, you can reach out to our team on 13 2221 or send us a message through the CommBank app. ^Mary.
1
1
1
This is my presentation at the linguistics conference. One hour including Q&A. I explain the use of context in conversation with examples.
0
1
2
RRG 2025 has wound up. Here's some nice images from the event of the scale of reference materials developing Role and Reference Grammar for linguistics use and therefore for exploitation in AI. My talk was on how and why RRG will be central to the Next Generation of AI, such as.
1
1
5
Understanding isn't just memorization. Nothing highlights the effectiveness of communications than being in another country and reading your own language written by a second-language speaker. The next generation of AI is coming from the cognitive sciences because our
2
2
7
I'll be talking at the RRG2025 event in Kobe, Japan on July 11 (link below). Topic, how to exploit linguistics in human conversations using the system I built from 2006. Also, to improve AI and human language learning. If anyone has a question here, I can respond to some during
1
0
3