[melbourne-pug] AI and ML idea

[Mike Dewhirst]

On 2/08/2018 7:57 PM, Tim Krins wrote:
> Unfortunately, I'm not sure that Python would be a silver bullet here, 
> or even any kind of bullet.
>
> I would expect sorting with computer vision is already very very 
> possible, but would be slower and more expensive than huge shredders, 
> magnets, water baths and other 'passive' sorting methods.
>
> The idea of creating building materials out of recycled plastic has 
> been around for a while - check out companies like ByFusion - which 
> may come into their own as China rejects recyclable plastic shipments.
>
> Not trying to bring the mood down - I would love to build a home out 
> of recycled plastic (as long as I could be very confident about its 
> structural stability over time).

At the end of the day waste has to be dealt with. There is just so much 
of it. Someone somewhere will figure out how to do it.

The sheer volume demands automation and that should attract entrepreneurs.

I reckon it will be an incremental process which gets more and more 
granular over time. Separation into specific compounds - like fuels - 
can be done chemically or biologically to some degree. Eventually the 
intractable waste will need to be either incinerated at high 
temperatures or encapsulated/vitrified and buried like nuclear waste.

OpenAI is written in Python and that tells me Python is the language 
which will drive a lot of ML/AI work in the foreseeable. It isn't a 
silver bullet for sure but if waste sorting is ever to be done by 
autonomous swarms of robots it will probably be part of the solution.

I just did a bit of googling and see this idea is old hat really. It is 
being done already.

Oh well ...

Cheers

Mike


>
> Tim
>
> On Thu, Aug 2, 2018 at 9:43 AM Mike Dewhirst <miked at dewhirst.com.au 
> <mailto:miked at dewhirst.com.au>> wrote:
>
>     On 2/08/2018 5:07 PM, Andrew Stuart wrote:
>>     I’d like to see the packaging industry ended - or at least stop making single use wrappers.  The core problem is not reusing waste, it’s that we manufacture an infinite amount of the stuff and every day the packaging industry (actually the garbage manufacturing industry) pours out never ending tonnage of more.  We’re drowning in garbage but still paying the packaging industry to keep making it.  Sorry that’s nothing to do with Python.
>     Andrew
>
>     You may be entirely correct. However, until we get to a point
>     where wrappers actually are obsolete we need to clean them up and
>     prevent them from polluting.
>
>     Robots can learn to clean up for us. That is what I think it has
>     to do with Python :)
>
>     Mike
>
>>     On 2 Aug 2018, at 4:42 pm, Mike Dewhirst<miked at dewhirst.com.au> <mailto:miked at dewhirst.com.au>  wrote:
>>
>>     Hi all!
>>
>>     I attended a dangerous goods meeting yesterday and most of the time was spent discussing waste.
>>
>>     Some contractors who can no longer ship waste to China now pay a few months rent in advance for a factory, stack it to the roof with waste and when no more can be squeezed in they disappear leaving the landlord to sort it out. Too often, spontaneous combustion works its magic first and the Fire Service has to respond.
>>
>>     The scientific consensus was that we need a war on waste.
>>
>>     I was reminded of that Mumbai documentary which showed hordes of children sorting through vast piles of waste extracting sellable stuff.
>>
>>     It occurs to me that all the consumption care in the world (today's world anyway) isn't going to cut the mustard. We need armies of robots which can learn to recognise recyclable material. They can be specialised, small, medium and large. They can have gas detectors, spectrum analysers and other sniffers and detectors to help.
>>
>>     Eventually I see waste mountains being broken down into more granularly unique substances and chemical processes used for the hard-to-extract stuff. In (unproven) theory some of the waste should be convertable into energy to recharge the robot batteries and power their collaboration networks.
>>
>>     Even if some of the recovered materials were not worth selling they could be stored until someone thought of a use for them. House-bricks?
>>
>>     I wonder if any of you inventive pyfolk think the world needs swarms of intelligent waste converters?
>>
>>     Its an idea but with a self-organising collective it might become a project!
>>
>>     Cheers
>>
>>     Mike
>>
>>
>>
>>     _______________________________________________
>>     melbourne-pug mailing list
>>     melbourne-pug at python.org <mailto:melbourne-pug at python.org>
>>     https://mail.python.org/mailman/listinfo/melbourne-pug
>>
>>     _______________________________________________
>>     melbourne-pug mailing list
>>     melbourne-pug at python.org <mailto:melbourne-pug at python.org>
>>     https://mail.python.org/mailman/listinfo/melbourne-pug
>
>     _______________________________________________
>     melbourne-pug mailing list
>     melbourne-pug at python.org <mailto:melbourne-pug at python.org>
>     https://mail.python.org/mailman/listinfo/melbourne-pug
>
>
>
> -- 
> Tim K
> Jersey, CI
>
>
> _______________________________________________
> melbourne-pug mailing list
> melbourne-pug at python.org
> https://mail.python.org/mailman/listinfo/melbourne-pug

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/melbourne-pug/attachments/20180805/600d8222/attachment-0001.html>