That’s been my experience with outsourcing as well. You get exactly what was in the specs, no more, no less. Don’t expect performant code, scalable code, or secure code unless you exactly specify those targets.
I’m one of the DevOps guys, so I’m not against it in any way. I’m just disappointed that for YEARS I had tried to make progress in this direction, and it wasn’t until we decided to outsource everything as fast as possible that this got off the ground. Now we can’t do it fast enough, purely to be able to get rid of our full time folks.
This is what we’re seeing so far. They work to to document, and woe be to the business/analysts that didn’t describe what they wanted accurately. In their eyes, that’s IT’s fault though - so will be interesting if we rethink things eventually (the oursourcing contract is very long, so not much hope of turning the ship around)
Eh, in some ways, but it’s not quite fair to call it “bog standard”. Some of its advances and success are due to its immense hardware, but some of its successes are also due to how the guys on the Watson team have architected it.
I mean, “rules based system” is a super generic term. Most such systems can’t achieve what Watson can. There’s nothing “standard” about Watson’s capabilities.
I’m almost 40. Imagine how I feel!
Really? My understanding is that there is no extant “Watson” technology; IBM just slaps the Watson name on any solution they are offering that has some rule-engine or ML element; the actual software they deploy has nothing to do with the (quite-sophisticated) NL/search engine they built for Jeopardy (except perhaps in those few cases where advanced NL capabilities are required). The actual rule and ML components deployed are based on industry-standard open-source platforms for those technologies.
I don’t have any first-hand experience, though, so I obviously could be wrong.
I do, and this is simplified, but not inaccurate. Many of their enterprise systems have ‘Watson’ in them, which really just means they name any remotely linked modules Watson for branding purposes.
The enterprise systems I’ve used at work all added this in the last few years, and it has only tenuous links to the original.
I think that part of what you are saying is correct, in that “Watson” is kind of a brand name now.
But at the same time, I personally know a guy who works on the AI Watson team. They’ve actually developed a system beyond just the brand name.
I was hanging out with my brother in law this weekend and he brought up automation and machines eliminating human positions in a field I hadn’t previously considered when thinking on this topic - law.
He’s a lawyer, and he brought up how a lot of the big firms are reducing the number of first year associates they bring in on an annual basis due to the sophistication of computers at performing a lot of the grunt work they were generally used for in the past - paperwork searches, collating relevant data, etc.
It was interesting to me primarily because I often think of automation taking over in areas such as manufacturing, retail, shipping, etc.
What about finance? Computers make stock market decisions much faster than humans.
Yeah, but someone has to get rich off that shell game.
Computers will always be better at information than physical tasks. Also the rate of improvement is much better.
@nijimeijer I would think if those types of tasks are the ones getting axed, maybe Paralegals would be on the chopping block too. I don’t exactly know what a Paralegal does but in my mind it’s “semi-lawyerly stuff” since “lawyerly stuff” would be done by those first-year associates.
I have had need for a lawyer twice in the last 10 years, both dealings with my parents trust. Paralegals did 90% of the work helping to reduce the cost.
I worked at a legal firm as a document processor. In my experience the paralegals did the research. Then the lawyers took that to court. The paralegals did 99 percent of the work. But the lawyers spoke in court.
Automation will make court costs so cheap, anyone will be able to sue anyone for anything! I look forward to this eventuality.
This is basically already happening:
(quick copy paste for the some things the site does, which is a legal bot service)
Parking Tickets
Reduce Property Tax
Sue Someone
Fight Your Landlord
Credit Score Errors
Reduce Rent
Get A Refund
Contract Disputes
Get Cash Back
Generate A Will
Compensation For Data Breach
Divorce
Immigration
Freedom Of Information Act Requests
Help With Bankruptcy
There’s an AI project that makes 2D platformers. It “looks” at a game, something like Mario for instance, and from that, it manages to code something that resembles that Mario game. Not as good, but recognizably Mario.
Still, I’m sure it’s far from a general case tool, and the tech to do a general use “Make me a word processor” doesn’t exist. Because if and when it does exist, “make me a better version of yourself” will end up with a much smarter than any human version in no time flat…
There’s not much more general purpose than a video game. Video games need GUIs, lots of math, narrative, architecture, etc. etc. Yet procedural generation in video games is still pretty lame, usually. And video games tend to be pretty close to the cutting edge, technologically.
Here’s a story on the bright side of job automation. An autonomous helicopter has flown an operational military flight to deliver supplies. I for one am looking forward to even greater automation in the military to get people off the front lines (or any lines).
robot repair robot repairers, yeah
“This was rare, candid and bracing talk from executives atop corporate America, made at a conference Thursday at the Dallas Fed. The message is that Americans should stop waiting for across-the-board pay hikes coinciding with higher corporate profit; to cash in, workers will need to shift to higher-skilled jobs that command more income.”