A lot of freaking out is about the A.I. automations of the future that can think like us, talk like us and know everything. However, I am doing regular computer automation with regular code that works with normal data. In the process transforming a company from working 1990's style 'personal computing' to where the web server 'does all that for you'. Many thousand of 'if' statements later I have automated many things have changed, three teams working from same-but-different spreadsheets now just use the same data in a proper database. If anyone enters the data wrongly then the computer corrects it and probably had it right before someone changed it.
The thing is that some quite 'simple' office jobs are quite complicated to do in code, debug and test. But then you do it 'the computer way' so it can check thousands of records or millions of them to see if that email address has been used before. A mere human would not know that or be able to just casually check a million or so previous sales orders etc. in a couple of seconds. So with these 'computer special' ways of doing things, the finished automation goes beyond what is possible by hand, or if attempted by hand the time taken would just not be available. The human role has been deskilled and the computer code is already approaching 'self awareness' even though it is just a lot of if statements with no Google brain A.I. in sight.
A lot of thought about information architecture is needed for this, more like applying the Toyota Way to 1's and 0's with changes to business processes and new properly thought out databases of sorts built to support this new 'web' way of working. Essentially lots of data entry work is taken out of all departments in the company with the web replacing documents as we know them.
If you ream through a company taking the data entry out of a person's job here and there, this is usually a help because it does free these people to be more productive and do the things that they wanted to do and joined the company for, which is rarely data entry. They can progress within their career and use new tools that are part of a better company wide team thing, enabled by the web.
Companies that do not automate like this cannot compete on price with those companies that do get 'web' ways of working properly automated. The companies that don't fully embrace the web tend to have 'the web' as a hybrid 'typing pool' of sorts, a cost, they don't get it, but it happens.
In this future the company that automates its processes just with if statements, cron jobs, a sprinkling of SQL in the reporting and everything going through https://, this company can scale. If all the customers decide to return product for a refund at the same time (e.g. Note7) then this can be handled without a bottleneck in back office processes and resultant customer dissatisfaction about money not returned etc. In this way automation is crucial for existential threats that could ruin a company's reputation. This scale-able company is able to take up the market occupied by non-automated rival companies, not visa-versa.
'If' statements are very cheap when written and tested, sending data is also very cheap. It is this simple, non-AI automation that is the thing people need to fear. Despite how many hours someone works or time served or how high up the hierarchy they are, there is the risk that half of their job could go, to be replaced by some helper module code that is a mere handful of kilobytes. Relatively speaking the Google grade AI self-driving car stuff that replaces rocket surgeons is not the thing to be scared of, it is the insidious 'if' statements of regular code and how the winners will all have to use 'web' ways of working.
The thing is that some quite 'simple' office jobs are quite complicated to do in code, debug and test. But then you do it 'the computer way' so it can check thousands of records or millions of them to see if that email address has been used before. A mere human would not know that or be able to just casually check a million or so previous sales orders etc. in a couple of seconds. So with these 'computer special' ways of doing things, the finished automation goes beyond what is possible by hand, or if attempted by hand the time taken would just not be available. The human role has been deskilled and the computer code is already approaching 'self awareness' even though it is just a lot of if statements with no Google brain A.I. in sight.
A lot of thought about information architecture is needed for this, more like applying the Toyota Way to 1's and 0's with changes to business processes and new properly thought out databases of sorts built to support this new 'web' way of working. Essentially lots of data entry work is taken out of all departments in the company with the web replacing documents as we know them.
If you ream through a company taking the data entry out of a person's job here and there, this is usually a help because it does free these people to be more productive and do the things that they wanted to do and joined the company for, which is rarely data entry. They can progress within their career and use new tools that are part of a better company wide team thing, enabled by the web.
Companies that do not automate like this cannot compete on price with those companies that do get 'web' ways of working properly automated. The companies that don't fully embrace the web tend to have 'the web' as a hybrid 'typing pool' of sorts, a cost, they don't get it, but it happens.
In this future the company that automates its processes just with if statements, cron jobs, a sprinkling of SQL in the reporting and everything going through https://, this company can scale. If all the customers decide to return product for a refund at the same time (e.g. Note7) then this can be handled without a bottleneck in back office processes and resultant customer dissatisfaction about money not returned etc. In this way automation is crucial for existential threats that could ruin a company's reputation. This scale-able company is able to take up the market occupied by non-automated rival companies, not visa-versa.
'If' statements are very cheap when written and tested, sending data is also very cheap. It is this simple, non-AI automation that is the thing people need to fear. Despite how many hours someone works or time served or how high up the hierarchy they are, there is the risk that half of their job could go, to be replaced by some helper module code that is a mere handful of kilobytes. Relatively speaking the Google grade AI self-driving car stuff that replaces rocket surgeons is not the thing to be scared of, it is the insidious 'if' statements of regular code and how the winners will all have to use 'web' ways of working.