Slashdot Mirror
Disruptive AI Bots Are Aleady Delivering Radical Leaps In Productivity (venturebeat.com)
The CTO of Textio is describing the "already happening" AI disruption that no one's noticed, arguing that voice-activated assistants are "just one small part of what AI is about -- and not the part that will matter the most for the enterprise companies that actually buy almost $4 trillion in software and services each year." An anonymous reader writes:
Jensen Harris describes "the less-flashy flavor of AI that is changing the nature of work itself: headless AI...the application of artificial intelligence to vastly improve internal business processes. It is fully transforming the crucial machinery of business -- processes like hiring, lead generation, financial modeling, and information security. Legacy software has become a commodity in all of these areas, and purpose-built AI solutions will get a larger and larger wallet share of these huge enterprise cost centers."
Combining machine intelligence with learning loops, these constantly-evolving algorithms are "where the money is," since headless AI "doesn't try to replace people; it gives them superpowers" -- for example, predicting the future. Harris ultimately argues that headless AI are delivering "radical productivity leaps that they haven't seen from software in decades... In the near future, every core business function will have been transformed by AI -- hiring, sales, security, marketing, finance, manufacturing...everything... Legacy software will get squeezed down into a smaller portion of the IT wallet as the most valuable services become the native AI platforms -- just as form-based desktop software got squeezed out by the cloud in the last generation... the real enterprise revolution is happening in the companies that are using headless AI to transform their core businesses."
By comparison, he argues that many of today's bots "are kind of a hipster facade around the same basic command line interfaces consumers abandoned in the 1980," and suggests this focus on personality misses the larger significance of behind-the-scenes AI.
Combining machine intelligence with learning loops, these constantly-evolving algorithms are "where the money is," since headless AI "doesn't try to replace people; it gives them superpowers" -- for example, predicting the future. Harris ultimately argues that headless AI are delivering "radical productivity leaps that they haven't seen from software in decades... In the near future, every core business function will have been transformed by AI -- hiring, sales, security, marketing, finance, manufacturing...everything... Legacy software will get squeezed down into a smaller portion of the IT wallet as the most valuable services become the native AI platforms -- just as form-based desktop software got squeezed out by the cloud in the last generation... the real enterprise revolution is happening in the companies that are using headless AI to transform their core businesses."
By comparison, he argues that many of today's bots "are kind of a hipster facade around the same basic command line interfaces consumers abandoned in the 1980," and suggests this focus on personality misses the larger significance of behind-the-scenes AI.
Sounds redundant. But basically true. Seriously though, there is very little substance in the latest "AI revolution", just like every other "AI revolution" in the last 70 years. The solutions still lack adequate knowledge of context and still push tricks and training depth beyond their effective limits.
Strange things are afoot at the Circle-K.
Hipster facade indeed.
Smelling some serious Silicon Valley marketing bullshit in here. Damn, the whole summary is a buzzword bingo paradise. Question: what AI *isn't* "headless"? He just means "non-customer facing", right? Don't overload the term in confusing ways.
Am I just grumpy this morning, or is this summary as asinine as it seems to me right now? Not only is this schmuck cranking the AI hype train whistle up to 11, he's not nearly as insightful as he thinks it is. Um, yeah, no kidding that AI (a fancy term for advanced data analysis, mostly) will be useful to businesses, instead of the stupid chat bots now being displayed. And yeah, it seems pretty obvious it's not going to put everyone out of work (at least not in the short term), but will just be another tool people use.
Maybe I'm just getting tired of the over-hyping of AI. Almost makes me miss the 3D printing fad.
Irony: Agile development has too much intertia to be abandoned now.
Can we have AI that will tell us how exaggerated these claims about AI are?
It's too bad we can't just get information on topics like this. It has to be amazing and dramatic!
Yes, all the functions mentioned benefit from ML algorithms. There are startups applying the whole range of stuff you'll find in MLlib, scikit-learn, etc. to these problem domains.
This is hardly article worthy and "headless AI" is pretty meaningless and lame as a catch phrase. Nothing new here.
When practitioners use the word AI, we know we are being a bit tongue in cheek with it. But most of what people in the industry would call (weak) AI is really conversational AI and other hard NLP/NLG problems, and computer vision domain problems.
That subsumes two of the primary applications of deep learning / deep neural networks. But those are just techniques that happen to be pretty good at feature learning with large data sets. Connectionist approaches may work better these days, but they dont even make for weak AI on their own, outside of toy scale problems (see Facebook BABI, for example).
In any case, every generation of computer scientists redefines AI to mean the really cool adaptive or statistical algorithmic approaches the last generation didn't have. So at one point Rete nets and rule engines were AI, then neural networks were AI, then SVMs and kernel tricks were AI, now deep learning is AI. But these are algorithms they aren't in and of themselves AI.
In my experience, effective weak AI requires staged learning - mapping unstructured data to structured representations and then processing combinations of more unstructured data with that structured representation to interpret and complete a real task in a human-like manner. All at once deep learning approaches work great for research papers but less well in practical use.
When the AI headless robots come, count me in with the robot smashers.
8 of 13 people found this answer helpful. Did you?
Why pay inflated salaries to humans when my new bot can do it all:
do { make_decision(random()); } while (!is_company_bankrupt()); collect(large_severance);
what a fucking ignorant obfuscated line of bullshit. he should be killed
Deserves whatever they get.
We as computer people have trouble understanding how valuable is failure and non-perfectness in human society.
With AI taking decisions instead of us, we voluntarily give control to a third party and lose our ability to learn from our mistakes in the process.
Guys, if one day you can't get a promotion anymore because you have Asperger's and an AI judges the only job you can do is type code, what kind of dreams and future is that going to be?
Don’t claim these AI bot’s don’t or won’t replace people. Remember the golden rule. “The Goal of business is to make money”. Period. Without that goal you have socialism.
If a company can replace IT Security Staff, Doctors, Lawyers,-what ever - you can darn well bet they will. It’s a fiduciary requirement of the CEO as he is beholden to the board of directors and stock holders to produce a return on investment.
Don’t be naïve.
you're replacing people. Maybe they'll find other work. Maybe their departments will grow a little. But if the change really is radical and that's not just hyperbole than the rest of the economy won't be able to keep up. That's what happened with the industrial revolution. 80 years of unemployment strife and poverty before there was enough new tech to start employing folks in mass again...
Hi! I make Firefox Plug-ins. Check 'em out @ https://addons.mozilla.org/en-US/firefox/addon/youtube-mp3-podcaster/
Who would have guessed that an article written by a vendor of AI software says not only that it gives you superpowers, but if you don't jump on the train now your business is totally f*cked.
Apparently, if you're just another talking head in a suit that needs to sound cool and in-touch, "disruptive" is now the goto buzz-word.
I wonder how long it will be before it sounds just as cliche as "Synergy", "Leverage", "Win-Win", and "Outside the box".
It isn't like the 'meat' in every organization hasn't been attempting to implement screamingly obvious increases in efficiency, for hundreds of years. So there will have to be a Headless Headless ( H^2 ) to filter the results from the Headless Bots:
H^2 Bot: "You must be new here."
H^2 Bot: "We have always done it this way"
H^2 Bot: "It will make the 'X' feel bad"
H^2 Bot: "That isn't your department"
H^2 Bot: "Their is a team working on that already"
H^2 Bot: "We tried that before, and it didn't work"
H^2 Bot: "Did you run it by the team?"
In other words, we'll need Pointy-Haired Bots.
And everybody will go paperless, right?
when i think of ai i think of automation and more to a "protocol bot" by descriptive name. its software coupled with mechanics that that knows when and what under what conditions to do something. that can go into any ai field to personality and speech to more industrialized applications but does not learn though it could have sensors to know that it is being damaged and or being damaged and thus in effect feel physical damage which is what this kind of pain in humans is. we could argue that emotional pain too is just a sensor of distress as well but im not sure how that would be useful in automation outside perhaps a show of sympathy in some kinds of bots.
Is that the new buzzword for analytics?
When the AI bot decides the future of me based on my past, does this not create a feedback loop? Suppose one has a couple of trying years due to family leave, changing job market, changing jobs. Does AI figure this out and "give one a break", or is this a quick and easy way to say "no discrimination here, the computer is color blind, including being blind to the struggles of humanity."
I've been working professionally as a developer for well over two decades now, both server side for many years, and also more recently in mobile development.
The current AI revolution is NOT A FAD. I repeat, not a fad.
The advancements in what neural networks can do, and do well has advanced in even just the last few years, Sure ten years ago you were right as to what AI or neural networks specifically could do well, would not amount to much. But not today, and certainly not tomorrow.
I think perhaps putting these advancements under the flag of AI is a bit wrong, it's really all about neural networks specifically because that is what has made rapid advancements and it already working really well on a variety of problems. If you are a developer today I think it is an absolute must to study neural networks now so that you can be on the advance of the train rather than run over by it.
In the end NN are just another tool a developer can use to build applications and systems. But understood and used well they can be even today a very powerful tool, and there are a lot of forces advancing this field rapidly to make this even more true in the near term...
"There is more worth loving than we have strength to love." - Brian Jay Stanley
I want Sandy was a nice start at an expert system. Sorta like procmail, with a little more.
As a corporation, we want a self-hosted solution. We are NOT interested in using external "AI" solutions which might leak.
I'd love a self-hosted cause-effect solution to control devices around the house. "Illuminate" "Volume Up" etc.
Maybe we could get some AI to filter out articles by people who self-promote their companies or products.
As we understand how people do work and optimize on that work using technology then yes the systems will start taking over that work. This has been happening since the steam engine. The work today is more complex, but selling items to people is the same and it can be optimized. The business that use technology in a different way to gain an advantage will win out over those that don't.
The current AI works for some things and not others, good a pattern recognition, but it will need to have a better higher level decision maker if it's going to do something at the creative human level.
I'm still waiting for someone to release box. It's in Star Cops.