Algorithmic Ouroboros...

[Fred Bloggs]

\"While most AI models are built on data made by humans, some companies are starting to use — or are trying to figure out how to use — data that was itself generated by AI.

\"Overall, the moonshot that companies like Cohere are working toward is self-teaching AIs that generate their own synthetic data.\"

https://futurism.com/the-byte/ai-synthetic-data

Sounds crazy until you find a sane source, rather than mass media hysteria. It\'s much more mundane in reality, not exactly a big leap:

https://research.ibm.com/blog/what-is-synthetic-data

 

[Phil Hobbs]

On 2023-07-22 11:29, Fred Bloggs wrote:

\"While most AI models are built on data made by humans, some companies are starting to use — or are trying to figure out how to use — data that was itself generated by AI.

\"Overall, the moonshot that companies like Cohere are working toward is self-teaching AIs that generate their own synthetic data.\"

https://futurism.com/the-byte/ai-synthetic-data

Sounds crazy until you find a sane source, rather than mass media hysteria. It\'s much more mundane in reality, not exactly a big leap:

https://research.ibm.com/blog/what-is-synthetic-data

\"Starting from \"I think therefore I am\", it had deduced the existence of
rice pudding and income tax before anyone managed to switch it off.\"
- Douglas Adams

Cheers

Phil Hobbs

 

[Jan Panteltje]

On a sunny day (Sat, 22 Jul 2023 08:29:24 -0700 (PDT)) it happened Fred Bloggs
<bloggs.fredbloggs.fred@gmail.com> wrote in
<5a658586-f7ae-4952-a604-e6101fb6e6b7n@googlegroups.com>:

\"While most AI models are built on data made by humans, some companies are =
starting to use — or are trying to figure out how to use — =
data that was itself generated by AI.

\"Overall, the moonshot that companies like Cohere are working toward is sel=
f-teaching AIs that generate their own synthetic data.\"

https://futurism.com/the-byte/ai-synthetic-data

Sounds crazy until you find a sane source, rather than mass media hysteria.=
It\'s much more mundane in reality, not exactly a big leap:

https://research.ibm.com/blog/what-is-synthetic-data

Self-teaching will take of when small AI powered robots start to do their own experiments.
That would be cool and also very dangerous, one [re]inventing nitro the
other one?
But the bots would be at it 24/7.
Better have them do it on some other planet.
Without practical experiments the self-teaching has no value.
Would get stuck in Einstein\'s relativity babble,
quantum stuff, build its own CERN...
Would take resources it could perhaps dig up for itself (using other bots).
Would it surpass us?

 

[John Larkin]

On Sat, 22 Jul 2023 12:58:45 -0400, Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2023-07-22 11:29, Fred Bloggs wrote:
\"While most AI models are built on data made by humans, some companies are starting to use — or are trying to figure out how to use — data that was itself generated by AI.

\"Overall, the moonshot that companies like Cohere are working toward is self-teaching AIs that generate their own synthetic data.\"

https://futurism.com/the-byte/ai-synthetic-data

Sounds crazy until you find a sane source, rather than mass media hysteria. It\'s much more mundane in reality, not exactly a big leap:

https://research.ibm.com/blog/what-is-synthetic-data


\"Starting from \"I think therefore I am\", it had deduced the existence of
rice pudding and income tax before anyone managed to switch it off.\"
- Douglas Adams

Cheers

Phil Hobbs

Just in time!

 

[Fred Bloggs]

On Saturday, July 22, 2023 at 1:09:26 PM UTC-4, Jan Panteltje wrote:

On a sunny day (Sat, 22 Jul 2023 08:29:24 -0700 (PDT)) it happened Fred Bloggs
bloggs.fred...@gmail.com> wrote in
5a658586-f7ae-4952...@googlegroups.com>:

\"While most AI models are built on data made by humans, some companies are =
starting to use — or are trying to figure out how to use — > >data that was itself generated by AI.

\"Overall, the moonshot that companies like Cohere are working toward is sel> >f-teaching AIs that generate their own synthetic data.\"

https://futurism.com/the-byte/ai-synthetic-data

Sounds crazy until you find a sane source, rather than mass media hysteria.> > It\'s much more mundane in reality, not exactly a big leap:

https://research.ibm.com/blog/what-is-synthetic-data
Self-teaching will take of when small AI powered robots start to do their own experiments.
That would be cool and also very dangerous, one [re]inventing nitro the
other one?
But the bots would be at it 24/7.
Better have them do it on some other planet.
Without practical experiments the self-teaching has no value.
Would get stuck in Einstein\'s relativity babble,
quantum stuff, build its own CERN...
Would take resources it could perhaps dig up for itself (using other bots).
Would it surpass us?

Just like human learning, the self-learning AI needs mentoring, probably by a set of inviolable rules it must follow, things like the reliability of the sources of its information. That\'s one reason they\'re saying the web is too \"noisy.\"

I don\'t know the first thing about it, not interested.

 

[Jan Panteltje]

On a sunny day (Sat, 22 Jul 2023 08:29:24 -0700 (PDT)) it happened Fred Bloggs
<bloggs.fredbloggs.fred@gmail.com> wrote in
<5a658586-f7ae-4952-a604-e6101fb6e6b7n@googlegroups.com>:

\"While most AI models are built on data made by humans, some companies are =
starting to use — or are trying to figure out how to use — =
data that was itself generated by AI.

\"Overall, the moonshot that companies like Cohere are working toward is sel=
f-teaching AIs that generate their own synthetic data.\"

https://futurism.com/the-byte/ai-synthetic-data

Sounds crazy until you find a sane source, rather than mass media hysteria.=
It\'s much more mundane in reality, not exactly a big leap:

https://research.ibm.com/blog/what-is-synthetic-data

Self-teaching will take of when small AI powered robots start to do their own experiments.
That would be cool and also very dangerous, one [re]inventing nitro the
other one?
But the bots would be at it 24/7.
Better have them do it on some other planet.
Without practical experiments the self-teaching has no value.
Would get stuck in Einstein\'s relativity babble,
quantum stuff, build its own CERN...
Would take resources it could perhaps dig up for itself (using other bots).
Would it surpass us?



PS
Like this for example:
New robot boosts solar energy research:
https://www.sciencedaily.com/releases/2023/07/230725122949.htm