Back in 97 Deep Blue (an IBM super-computer) beat the world’s greatest chess player…
At the time it was a seismic event that shook peoples understanding of technology against humanity.
Today, it’s not seen as anything special. Why wouldn’t a computer beat the world chess champion?
Garry Kasparov recently joked. “Chess apps on smartphones are far better than Deep Blue was.”
Today chess software developers hold their own tournaments & humans are excluded, as they wouldn’t stand a chance.
In recent years, the best chess software by miles was Stockfish. Until 2018! When Google company DeepMind turned up with some AI software called AlphaZero.
This is where the AI stuff gets scary…
AlphaZero is not chess software & didn’t even know how to play chess.
Yet AlphaZero kicked Stockfish’s virtual butt.
Out of 100 games AlphaZero either won or drew every single game, not losing any.
Only last year AlphaZero crushed a new improved Stockfish yet again, over a 1,000 game marathon.
That’s the power of self-learning. Slowly, but surely AI will become better at everything.
I even read there’s an AI therapist (called Ellie).
She’s been created to identify signs of depression, anxiety & PTSD in US soldiers.
It seems the soldiers prefer talking to Ellie than a real person.
The day all this could become a problem, is the day an AI can write AI software better than we can.
When Tesla & SpaceX leader Elon Musk was asked ‘what are the most immediate threats to humanity’, he said 3:
‘Nuclear war & climate change’ then fell silent looking at his shoes in deep thought, before smiling & saying, ‘I just hope the computers decide to be nice to us.’
It’s not just Musk who’s worried how rapidly AI is developing, and how underprepared we are for its repercussions, other leading thinkers & scientists are feeling it too.
Yes this is big. But it’s like with any new technological development. It can bring good & bad things.
Certainly, we can get AI to help with the office environment, with many of the mundane repetitive tasks taken over by AI software, with an AI controlled robot even doing the scanning.
But how far do we want to go with this technology?
Shall we write programs for the computers & robots to operate all our businesses, selling all the stuff they make to each other & then share the profits out evenly with us humans?
Sounds great. I’m happy to give AI control over our office technology solutions business, just as long:
- The computers are ethical with each other
- They don’t hoodwink other computers
- AI upholds our fixed price guarantees
- No extra charges are brought in, through self-learning
- They don’t stitch up other computer brands, like IBM or HP, on long term contracts
- AI doesn’t make it complicated & they deliver Simple Honest Service
With AI, it really depends if we can all agree & stick to a set of global control measures.
Because, if we’re not careful. AI will reach a point where its intelligence will outstrip ours by so much, we’ll no longer have any control or idea what it’s doing.
…and they’ll be the 1’s lying on a sun-drenched beach in the Maldives drinking Pina Colada’s.
