This needs a cool name. There’s “The Singularity” – how about “The Feedback Loop” or perhaps “The Oroboros Loop”?
The Feedback Loop, AKA, CASE Oroboros Red, is when the humans are completely forced out of cyberspace by armies of robots chatting to each other. There are two aspects of this that will appear odd: one, the robots will rate-limit themselves to human speeds as they try to make a vague attempt to appear human and not gobble all the bandwidth – two, the robots will talk a great deal about penis enlargement and they have none. Don’t worry, some cyberbro will implement an E-penis and put it on the net, and then robo-dialers will somehow be able to use it.
All joking aside, we are getting to the I-don’t-give-a-fuck-ocalypse; the point where it all turns to self-referential garbage that’s not worth paying attention to, anymore. So, you already know that there are search engines, which score sites using complicated algorithms (Artificial Intelligence!) to decide what is interesting or not. The search engines spider the web, assess what’s interesting, and rank it – they also assess humans’ interactions with pages to determine what is interesting. Except, we all know that a significant percentage of “web hits” are robots. In other words, there are vast numbers of sites that are being read by robots. It’s OK. Apparently some of my best friends are robots and they are trustworthy. That’s what they tell me when they call on my telephone.
I imagine there are entire sites (like Twitter) that have no actual users; they’re just robots. And they post comments for other robots, which are programmed to respond. Sometimes the conversation grows heated, but that’s just a byproduct of statistically similar outputs being generated by well-trained commenter-bots.
Do I portray the future too darkly? I think I’m an optimist. It may mean that eventually we can stop playing the game entirely. There is only one problem: the marketers will follow us wherever we go. Every interesting new site, service, game, concept, communications medium – it will eventually fill with nothing but commenter-bots yattering at each other. Because they are infinitely patient, their victory is inevitable: their important message will get through. The only people who will not hear it will be those few who have decided to live without smart phones, internet, social media, constantly shifting to the next new platform as fast as the bots take over the old ones.
Now, we don’t need web developers to develop pages for robots to scan! We can have a robot web developer! [reg]
Microsoft has introduced an AI-infused web design tool called Sketch2Code that converts hand-drawn webpage mockups into functional HTML markup.
I strongly suspect that if you gave the AI a “sketch” that was actually a selfie of your smiling face, it would bash it into a perfectly WordPress-looking website. If you gave it a picture of your cat, it would probably produce another WordPress-looking site. I am absolutely sure that if you just rolled a couple of dice and used them to pick a WordPress style at random, you’d wind up with a better site. In fact, someone needs to implement an “I feel lucky” WordPress style template that randomizes itself every morning.
Microsoft, keen to coax customers toward its AI-oriented Azure services and to flaunt its data-processing prowess, would have web designers trade keyboard for pencil and let clever code interpret designer intent from doodles instead of relying on some app to slavishly respond to explicit commands.
Front-end web jockeys, freed from the burden of applying their expertise, can look forward to the creative satisfaction of quality assurance, a phrase which here means checking the AI’s work.
It might be argued that Sketch2Code lowers the web design bar – not all that high to begin with given the capabilities of existing tools – to admit casual scribblers into creative crowd while decreasing the time from whiteboard reverie to working prototype.
AIs could replace marketing directors. Uh-oh, I bet the marketing directors never saw that coming.
“Once you have drawn these wireframes on a whiteboard, you can take a picture using the web app and then the web app would send that information to the AI service,” said Tara Shankar Jana, senior product manager at Microsoft AI, while evangelizing the tech in a video. “The AI service then runs those images against the prebuilt AI model and then creates an HTML codebase followed by a resulting app.”
I ran that through my AI translator and it said:
“Once you have taken some random pictures of your cat, you can give the picture to a web app, which will send the information the AI service,” said Surly Badger, senior grumpus at Badgeria Brewery, while evangelizing the tech in a video. “The AI service then runs those images against prebuilt AI models of marketing campaigns and produces a complete marketing campaign with collateral. It even calls a publicist with a robo-call center and tries to pitch them an article about the new campaign.”
Meanwhile, across town:
A publicist is installing a chat-bot that will answer their phone line and interact politely with anything that calls.
The ad campaign that runs resembles something written by Alfred Jarry, but no human ever reads it.
This posting is dedicated to the memory of Authenticity. We hardly knew ye.
By the way, if I had a kid I’d name the poor unfortunate “Authenticity.” I guess we’re all glad I had a urologist take care of that.
Lofty says
I look forward to Stderr featuring robotic lolcat posts.
avalus says
First: @1: Exactly!
Second:
It reminds me of … a lot SMBC-Comics with the punchline: There are no humans! Its just Robots talking to Robots!
Third:
“Microsoft, keen to coax customers toward its AI-oriented Azure services and to flaunt its data-processing prowess, would have web designers trade keyboard for pencil and let clever code interpret designer intent from doodles instead of relying on some app to slavishly respond to explicit commands.”
Reminds me of the glorious self-praise my preferred Chem-Drawing program shouted at me to try to get me to use their ‘chemical interpretation mode’ for faster Structure drawing and faster database-searches. Only got total garbage I when I tried. “Ya typed ‘N’? Well, ya surely mean Na in this molecule. And Nd. And lets toss Nb in the mix as well!”
Why would you give up the use of explicit control, ever?
Dunc says
Never has there been a profession so thoroughly dedicated to its own elimination…
At least 10% of my Instagram followers are really obvious bots. I consider that a good percentage – I reckon they probably account for something like 90% of total activity. Robots taking to robots indeed…
cvoinescu says
the creative satisfaction of quality assurance
I can’t for one second imagine someone saying that with a straight face. But if they did, it should be up there with War is Peace, Freedom is Slavery, and Ignorance is Strength.
Cat Mara says
Ubu CIO?
Marcus Ranum says
Cat Mara@#5:
Ubu CIO?
By my Sword of Staff Meetings!
Intransitive says
Aaaaand we’re back to self-playing Pong games.
https://scratch.mit.edu/projects/72950668/
Nolan Bushnell would have been proud. Or apalled.
xohjoh2n says
In a previous life (just pre-dot-com-implosion) a certain consultant raised a super-important feature request that involved asking the user for some input to make a basic choice about something-or-other. Pretty simple. They could have just left it at that. But no, they had to specify exactly what the user-interface would look like. The requirements specification therefore was a hand-drawn Windows dialog box drawn on a napkin showing some random related window title, a smiley face in the middle of the dialog, and an OK box to accept the decision. “It has to look like that” they said.
Being the fastest expertest Windows dev (linux these days, much happier) I got roped into helping the lead implement the first revision of such a complicated feature – they had a specific idea about how to best to satisfy the consultant’s desires which I agreed would be the best initial approach.
The first revision was of course a dialog box that replaced the normal Windows dialog representation with a scanned image of the requirements image, with a normal OK button overlaid in the appropriate place. (By doubling the development effort from 30 minutes to an hour I could have even made the button look like the napkin version.) It met 100% their stated requirements, yet for some reason they weren’t happy with it and said it didn’t…
Ieva Skrebele says
Hahaha, that’s a good one.
Laughing aside, this is actually a serious problem for graphic designers. We experience such crap on a daily basis. Clients who don’t know a thing about art attempt to give us extremely specific instructions about how the design they commissioned ought to look like. In my case, it usually goes approximately like this: a client starts telling me what composition or color scheme or whatever else they want. In my mind I can visualize what they are asking me to make, and I already know that the design will look pathetic if I make it exactly as asked. The problem is—unlike me, my client cannot visualize the finished artwork, they cannot tell that what they want isn’t going to look good. At this point I have to try my best to say “your idea is stupid, if I make it exactly as you requested, the finished design will look like shit” in some manner that sounds polite and doesn’t offend my client. I work as a self-employed freelance designer. If some client refuses to listen to me at this point, I simply refuse to work for them.
It’s a whole different problem for graphic designers who work for some companies. I have an acquaintance who worked for a printing house. They had a client who was a lawyer and who needed a plaque for his office. He didn’t need anything complicated, just a basic plaque with his name and logotype. Such a job shouldn’t take a graphic designer more than about 10 minutes to finish. Except that this client knew exactly what composition he wanted for the plaque. My acquaintance made it exactly as requested, it looked ugly, the client realized that it looks bad and requested to redo it and make a different composition (once again, the client gave very specific instructions). My acquaintance did as requested, and once again it looked awful. The client asked for another revision. My acquaintance had to make over 10 revisions, each next one even uglier than what he did before. What should have been a 10 minute job ended up with him wasting hours.
By the way, I have worked as a designer for years, and so far only once I have experienced a case where my client gave me very precise instructions, and the end result was me being able to make an amazingly good looking design. The catch—my client himself was a designer, and he commissioned me to draw a tattoo for him. This time my client actually understood what he was asking. His instructions were clear, comprehensible, and the composition he asked for was doable and looked nice. That was this one https://www.deviantart.com/avestra/art/Cat-Head-Tribal-Tattoo-546085559 design I made.
For all other people who don’t understand design so well, it’s better to just let the artist/designed pick the composition, color scheme and any other details.
Johnny Vector says
Ieva Skrebele @ #9:
It’s Dunning-Kruger in action! I do know something about graphics and colors; I have won awards for lighting design, and been asked a couple times to light shows by a company I consider one of the greatest theater troupes in the country. And yet… I just finished a large proposal where we had a dedicated graphic artist. All my interactions with her were of the form “Here’s what I came up with. These connections need to be here, all of these objects should be tied together by color or something, and if there’s any way to make it fit on one page that would be great. Otherwise, just make it look good.” There were still multiple iterations in general, but y’know, why even have an expert around if you’re not going to use their expertise?