Re: DeepSeek

On Tue, 28 Jan 2025 0:25:21 +0000, Physfitfreak wrote:

On 1/27/25 5:52 PM, The Starmaker wrote:

Physfitfreak wrote:

>
On 1/27/25 1:55 PM, The Starmaker wrote:

I just started testing deepseek...
>
and it passed with flying colors!
>
>
Next post will show you my test and results...
>

>
You could not do that with chatgpt?

>
it's a comparison test...deepseek answers were more comprehensive.
>
>

>
>
Then OpenAI has tampered with normal functioning of Chatgpt to disable
that capability as well as fooling the users.
>
I've said it more than many times here that these tools and generally
science is in the hands of the wrong people. Cro-magnons cannot handle
it right.
>
Science will go where it belongs!

Science, in general, has been sequestered for a century or more by a
bunch of megacorporations funded, mostly, with public money in order to
accumulate power, wealth and with the main goal to privatize knowledge
in the hand of a few.
I imagine a disruption of this trend when hundred of spinoffs start to
emerge in the next years, making possible (working offline for security)
that almost anyone, with enough know-how and cheap resources, can
develop its AI engine "tuned" to work on any particular area, not only
in science but in any other conceivable field.
For instance, a particularly "tuned" AI engine to find patterns,
guidance and solutions in (random list):
- Stocks
- Biology
- Metallurgy
- Political engineering
- Profiling and exposure of fraudsters in any field
- Autonomous teaching machines, in ANY field (far from academia and
labs)
- Medicine, in every conceivable area. Capability for diagnosis free of
charge
- Working in the pharmaceutical area, tracing effects of medications,
etc.
- Analysis of millions of documents in arxiv, vixra and similar sites,
looking
  for patterns by correlating, interpolating and extrapolating.
- Detecting patterns of "online organized crime", from pedophilia to
drugs and
  fraud and corruption.
- Specialization in any kind of engineering and technology.
- Many other things, like watching for patterns of fraud and degeneracy
online.
And the above, plus an almost infinite list of areas, being able to be
explored FAR from TPTB.
It's almost too good to become true, but can change the future in a few
decades.
I'm sorry I'll not be able to watch it while it develops in the long
term.
Just imagine AI being used to trace corruption and crime, having partial
access to banking and fiscal information, as well as linking business
with illegal activities to accountable individuals.
What if, well provided with data, AI analyze the behavior of Boeing
since 1990?
But I'm digressing and dreaming.