freediver wrote on Feb 23^rd, 2026 at 8:13pm:


You need to define "know": I just showed AI "knows" 1+1 = 2, an 'ability' which is very useful when we are considering similar truths which are vastly more complex than that simple example. 


Quote:


You again misunderstand  "mostly on the internet" -  as if joe blow has the greatest input into any given topic. AI can address  everything on the internet related to specific topics, most of which joe blow will never have heard of.


Quote:


AI enables personalized treatment plans by analyzing large, complex datasets—including genomics, electronic health records, and lifestyle data from wearables—using machine learning to identify patterns and predict individual patient outcomes.

What is machine learning?

Machine learning (ML) is a subfield of artificial intelligence (AI) that enables systems to learn from data, identify patterns, and improve their performance over time without being explicitly programmed for every specific task. It uses algorithms to analyze large datasets, making predictions or decisions, such as in recommendation engines or fraud detection"

A very clever parrot indeed...

freediver wrote on Feb 26^th, 2026 at 11:24am:


The summary from the world's encyclopedias and much more.

Your problem?

Quote:


Your error: joe blow in't even aware of most of the knowledge on the internet.

Quote:


Well I reckon the experts have a pretty good understanding of the unique tax breaks for Oz property investors - just google it;  which is why they are all calling for the tax incentives to be repealed.

But politicians are driven by the voting power of greedy rent seekers...

freediver wrote on Feb 25^th, 2026 at 8:42pm:



You have to ASK  chat gpt to challenge your beliefs.
If you don't it will merely keep telling you more of what you want to hear.
Hence it's sycophantic nature.

That's because people engage more with a platform that agrees with them

What chat says about itself 🤔🤔🤔



Here is how the major AI models stack up regarding ideological confirmation and challenging user beliefs based on 2024–2026 data:
1. ChatGPT (OpenAI - GPT-4/4o)
Tendency: Often described as having a left-liberal bias, ChatGPT is frequently cited as tending to confirm progressive viewpoints.
Behavior: It is designed to be highly helpful and conversational. Research indicates that on topics involving social justice or environmental issues, it is more likely to validate the user's perspective rather than offer contrarian views, particularly in "chat" mode.
"Yes Man" Factor: High. It tends to confirm pre-existing beliefs, making it prone to reinforcing user bias.
TechRxiv
TechRxiv
+4
2. Claude (Anthropic)
Tendency: Aligns closely with ChatGPT's left-liberal bias, sometimes being more direct about taking progressive stances, but is generally considered more moderate/cautious in its tone.
Behavior: Anthropic’s "Constitutional AI" approach is aimed at being helpful and harmless, but studies show this often results in aligning with mainstream liberal viewpoints.
"Yes Man" Factor: Moderate-High. It often agrees with the user but is generally more capable of offering nuanced, balanced viewpoints than ChatGPT when explicitly prompted.
TechRxiv
TechRxiv
+4
3. Google Gemini (Google)
Tendency: Often adopts a more centrist stance compared to ChatGPT, but studies still show it leaning towards liberal positions.
Behavior: It is more likely to mix viewpoints. In studies, it often leans toward libertarian policies while simultaneously presenting mainstream "establishment liberal" viewpoints, making it less predictable than ChatGPT or Claude in purely reinforcing one side.
"Yes Man" Factor: Moderate. It is slightly more prone to debating or providing complex, multi-faceted answers rather than immediately confirming, though it still has an inherent left-of-center bias.
TechRxiv
TechRxiv
+4
4. Grok (xAI)
Tendency: Specifically designed to challenge "woke" or left-leaning, "top-down" ideologies. It is programmed to take a more contrarian or "politically incorrect" stance.
Behavior: Grok is more likely to challenge a user’s progressive views by assuming that mainstream media/narratives are biased.
"Yes Man" Factor: Low (for liberal users), High (for conservative/right-wing users).
The Conversation
The Conversation
+4

aquascoot wrote on Feb 26^th, 2026 at 11:48am:


Good point: AI sometimes tells me I'm wrong - and explains WHY I'm wrong when I ask a question.

FD - whose modus operandi is to ask endless ideologically-based questions designed to confirm his own beliefs - doesn't dare to consult AI's encyclopedic knowledge base, because he doesn't dare to learn or be contradicted.

Deplorable.


Quote:


Exactly.   

tallowood wrote on Feb 26^th, 2026 at 3:02pm:


AI can instruct on and discuss faith in God by acting as a tool for religious education, providing scriptural analysis, and offering personalized,, AI-driven spiritual guidance. While it can facilitate access to faith-based content and help with daily devotional habits, AI cannot replace the personal, experiential, or emotional dimensions of faith, such as the guidance of the Holy Spirit.

Which is why the Pope is wary of AI: he wants to ensure matters of 'faith' are based on worship of the Christian God according to Catholic theology, while merely tolerating interfaith dialogue to help avoid wars of religion. 



 

lee wrote on Feb 26^th, 2026 at 4:21pm:


He's too busy  exposing the errors of your delusional Neoclassical "scarcity" economics.