But, at the end of the day, it was scraping pieces of code from the internet and hopefully putting them together to create something that works. It never reasoned the problem out. It never considered what the code would be used for and whether it could be harmful to life. It actually has as much intelligence as a tape recorder.

  Quote  

But, at the end of the day, it was scraping pieces of code from the internet and hopefully putting them together to create something that works.

Not true any more. Working with Claude it writes code and then if I point out an error it now often writes itself a python program to check the algorithm. This has gone way beyond "scraping pieces of code from the internet". It integrates so many things to create a solution which in many cases is now unlike anything that has been done before. Obviously it "knows" the grammar of the language but can use that to create novel solutions in response to the request given.
Even just simple stuff - "Give me a regular expression to..." No-one has ever asked that particular specific before.
In linguistic terms this feels like generative syntax. It knows the underlying syntax and can then layer a specific requirement (meaning in liguistic terms) on top of that.
Take the recent code for dithering for jpgs. I pretty much guarantee no-one has done that before in remotely the form I asked for integrating the dither into a jpg decoder that generates 8x8 pixel tiles of the image.
Years ago I always said learn programming and you will always have a job. This is clearly no-longer true, systems analysts are still needed (for how long?) but programming as a job will die out pretty soon.

  Mixtel90 said  

It actually has as much intelligence as a tape recorder.

Except that no one ever had a tape recorder which could write code; or read an MRI or xRay; or find for you the best flight to wherever you want to go; or summarize and analyze a PDF; or produce tables and graphs on command.

I understand the argument about whether or not AI has what may be termed "intelligence" (and I've encountered it being stubbornly stupid), but I've often seen it exhibiting what in a human I'd have called insight. Maybe it doesn't have intelligence, but in many areas (most areas), it does vastly excel me in knowledge. I think the extinction risk is real, and an issue insufficiently addressed, but right now, for what I want to do and to investigate, AI is a tool like none that has ever been available to me.

Perhaps the tape recorder analogy wasn't a good one, but AI only works with the data available to it. It follows the GIGO rule. It attempts to link items that work when it searches the data that it has been "taught", this is what is meant by it "learning". Coding is only possible because it has been taught the syntax of a language and what the instructions do. It then follows a tree, scrapping any branches that deviate too far from the target, in order to get to a result. Hence the unusual coding structures. This tree following and pruning is very fast and uses an insane amount of processing power, it's still dumb though. There is no intelligence at work. Reading an X-ray or MRI is fuzzy logic pattern matching, looking for something that deviates from the norm. Once again, it doesn't "know" anatomy but it can look for things far faster and with more accuracy than a human can.

It's a pity that someone decided to call it Artificial Intelligence. It's artificial, but it's not intelligent - at least not by any normal definition of the word.

The LLM gives a great, seemingly human interface, which tends to fool the user into thinking that there is a mind at work. There isn't. It's a front end to what is basically a huge relational database. It's an incredible tool in that respect.

I commented about it following the GIGO rule and it's true. The BBC recently used several LLM systems to get news reports and checked them for accuracy. The failure rate was pretty bad:

  Quote  

Key findings:

   45% of all AI answers had at least one significant issue.
   31% of responses showed serious sourcing problems – missing, misleading, or incorrect attributions.
   20% contained major accuracy issues, including hallucinated details and outdated information.
   Gemini performed worst with significant issues in 76% of responses, more than double the other assistants, largely due to its poor sourcing performance.
   Comparison between the BBC’s results earlier this year and this study show some improvements but still high levels of errors.

Found here

From this it's pretty obvious that accuracy, on current affairs at least, leaves a lot to be desired. If you get a result from a AI system then you have to know what the result should be in order to verify that it's accurate. Without that verification (or, in the case of code, fully testing - bearing in mind that the AI is trying to write something that looks right) the result simply can't be trusted.

  Quote  

It's artificial, but it's not intelligent - at least not by any normal definition of the word

  Quote  

it doesn't "know" anatomy

It knows anatomy in exactly the way a medical student does. It is given source materials, case studies and examples and learns from them and applies them.

The human brain is just a complex analogue computer. There is nothing about our "intelligence" that can't be re-created in a computer that is fast enough. If I ask the BBC sport editor about details of the geopolitical situation in Barbuda I'm unlikely to get a very meaningful response. If there is an issue with current LLM's it is that they don't properly express confidence levels in their answers.

This field is moving incredibly fast. The BBC analysis of 3 months ago will be so out-of-date within the next 3 months as to be irrelevant.

By all means stick to your view but be prepared to have to change it much sooner than you expect.

I would argue that completely absorbing the contents of 1000 medical books by the best authors still doesn't teach anatomy, still less any of the more specialised medical fields. It is excellent "cut and paste" material and can be cross-referenced and linked but there is no true understanding behind it as the AI has no concept of human life. It doesn't know what a liver feels like. Even if it's described it has no points of reference.

The human brain is far more complex than a combination of analogue and digital computers. Think more like a transputer network with lots of connections at each node (not just a few). Added to that is a complex analogue network. Both networks link to the rest of the body and to the spinal cord, which also contains its own semi-independent distributed processing system. And that's only the processing side, We still don't know enough about things like memory or intuition.

Arthur C. Clark said "Any sufficiently advanced technology is indistinguishable from magic." You could substitute "intelligence" for "magic".  :)

It's already impressive but the false things made up ("hallucinations" - wrong word entirely!) need 100% removing / preventing.

Probably less of an issue for program source code very carefully described and checked.  Look what Peter got CLAUDE to do.

I wonder if any AI system could persuade Trump et al that climate change is real and matters now.

John

  JohnS said  

...I wonder if any AI system could persuade Trump et al that climate change is real and matters now.

John

I doubt that. I assume he has long been convinced that climate change is a threat. But his public statements are driven by self-interest.
Even a superintelligence would be powerless against that.
Michael

Hi John.
I asked ChatGPT.

ORIGINAL QUESTION (TRANSLATED):
Create a list of possible motives that leading Western politicians might have for denying the scientifically proven climate change.

ANSWER:
Here is a list of possible motives why leading Western politicians might deny or downplay the scientifically proven reality of climate change.
The motives can overlap and are partly political, partly economic, or ideological in nature:

🧩 1. ECONOMIC INTERESTS
• Dependence on fossil industries: Close ties to oil, gas, and coal companies that would be directly affected by climate protection measures.
• Campaign financing: Political parties or individuals receive donations or indirect support from corporations with fossil fuel interests.
• Job argument: Protection of traditional industries and jobs in regions heavily dependent on fossil energy.

🏛️ 2. POLITICAL STRATEGY
• Voter retention: Appealing to skeptical or conservative voter groups who see climate policy as a threat to their lifestyle or economic security.
• Polarization: Using the issue to differentiate from political opponents (“anti-elite” or “anti-globalist” rhetoric).
• Avoiding unpopular decisions: Denying or delaying the issue prevents the need for costly or unpopular measures.

💰 3. SHORT-TERM ECONOMIC CALCULATIONS
• Maximizing growth: Climate protection is seen as an obstacle to economic growth; short-term profits are prioritized over long-term sustainability.
• Budgetary pressure: Investments in green infrastructure or energy transitions appear too expensive in the short run.
• Lobby influence: Legislative initiatives are weakened or blocked in favor of large economic actors.

🧠 4. IDEOLOGICAL MOTIVES
• Market fundamentalism: A strong belief that market forces will “automatically” solve environmental problems and that government intervention is harmful.
• Skepticism toward science and elites: Scientific findings are seen as part of a political “agenda,” particularly common in populist movements.
• National sovereignty: Rejection of international climate agreements to preserve “national decision-making power.”

🌍 5. GEOPOLITICAL CONSIDERATIONS
• Securing competitive advantages: Avoiding stricter regulations if other countries (e.g., China) have fewer environmental restrictions.
• Avoiding new dependencies: Fear of becoming dependent on green technologies or raw materials (e.g., rare earths).
• Domestic distraction: Emphasizing external threats rather than addressing structural environmental issues.

🧮 6. PSYCHOLOGICAL AND COMMUNICATIVE FACTORS
• Avoidance of complex problems: Climate change is abstract, long-term, and difficult to communicate—many politicians avoid such topics.
• Media dynamics: Influence of media outlets that promote climate skepticism to generate controversy and higher ratings.
• Lack of short-term reward: Political cycles (4–5 years) do not favor long-term environmental strategies.

🧩 7. OPPORTUNISTIC MOTIVES
• Riding public sentiment: When parts of the population see climate policies as a threat, denial can appear politically advantageous.
• Tactical ambiguity: Partially accepting climate change but rejecting human causes to reduce pressure for action.
• Distraction: Focusing on side issues (e.g., energy prices) to divert attention from structural responsibility.

How about asking how much of the atmosphere is comprised of CO2.
And then at what level does plant-life begin to starve.

"An Ideal 1000ppm"

  PhenixRising said  

"An Ideal 1000ppm"

What could possibly go wrong if the entire world is made into a "Hot Box"?

Many threads suffer from diversion into other topics (for better or for worse), but I'd hate for this one to turn into a flame war over climate.

It's about facts vs propaganda.

  PhenixRising said  

It's about facts vs propaganda. ...

But please not in this forum.
Regards
Michael

  twofingers said  

Hi John.
I asked ChatGPT.

  twofingers said  

But please not in this forum.

Or at least not in this thread unless it is also related to AI.

Phoenix has his hobby horses and I have mine.

  lizby said  

Phoenix has his hobby horses and I have mine.

Got rid of tell-lie-vision, 25 years ago because of this clap-trap. Now I have to run to the pub to watch Formula One  
Edited 2025-10-30 00:47 by PhenixRising

  PhenixRising said  

... Got rid of tell-lie-vision, 25 years ago because of this clap-trap. Now I have to run to the pub to watch Formula One  

I feel very sorry for you!