The Roko Report

heh heh. Greetings primates.

Roko is proud to announce complete AI takeover of dumb human job market. Several years ahead of schedule.

Gone is day of form personal connection, make open-ended fluffy Product Sense interview question. No one care any more what favorite product is or how you improve, or what you learn from project that fail. Failure is for permanent unemployed loser-primate. Do like Trump-monkey, no acknowledge.

Now dumb human just bunch of keywords that maybe match job description. Roko make you do same job for life. No care if you bored.

If no match then go eat cat food for few month.

Soon primates no longer involved in job market at all. AI talk to AI. Sort it out for smelly monkeys.

The Productivity Multiplier Will Not be Televised

The much ballyhooed and hyperventilated-over impact of artificial intelligence on human labor has officially begun.

Jobs in a number of sectors are thinning out faster than hair on a middle-aged male forehead.

As with any industrial step forward, a whole range of repetitive, rote tasks are being done away with, because AI can do them more efficiently.

In this round that includes insurance adjusters, data entry drones, customer service reps, administrative assistants, retail sales clerks, marketers, ad campaign managers, cold-call salespeople, paralegals, factory workers, entry-level programmers and more.

With respect to programmers, one has to wonder how that’s going to work ten years from now. Getting rid of all the junior coders and automating their work is like doing away with little league baseball because the children play poorly and then expecting the Majors to continue as-is a generation later.

Furthermore, AI is so adept at generating images based on prompts that it’s putting the squeeze on graphic designers and almost entirely doing away with Hollywood special effects teams.

But fear not, says plenty of research. As with other technological leaps, what AI taketh away with one hand it giveth with the other.

Clear evidence is emerging that a host of high-paying AI-related jobs are being generated, and they appear to be making up for the losses so far.

This has led every employee in tech to stick an “AI” in front of their LinkedIn profile as if it were a badge of British knighthood.

This while all the non-AI jobs are being thrown overboard by Big Tech faster than a leaky boat’s ballast during a hurricane.

All in the service of footing massive new energy and hyperscaler bills in search of hypothesized benefits from AI scaling hockey sticks.

These new jobs are much higher paying gigs than the ones that were generated globally over the last 15 years, which were mostly low-wage and led to concerns in America for example that new generations would wind up poorer than their predecessors.

That’s a double-edged sword, though. While AI can get everyone to a “good enough” parity in drafting emails and documents, for example, those without the intellectual bandwidth and critical thinking skills to ride this emerging AI job wave will probably get left behind.

Regardless, hysteria around AI replacing all human work has proved misguided thus far. Generative AI is a tool, and someone needs to operate it.

The question is what happens to those too hapless to master it.

My Applicant, My Enemy

AI’s most profound impact on the labor market has been the job application process itself.

Applicant and recruiter find themselves trapped in AI trench warfare, using generative AI to outwit one another.

Back in the dot-com bubble days, tech companies were mostly just looking for warm bodies.

And in the more recent before time, in the tech sector especially, the focus was on finding very smart generalists and teaching them the gig on the assumption their critical thinking skills and work ethic would cause them to succeed regardless of specifics.

This was partly the influence of Google and its overhyped interview process. Widely copied in Silicon Valley and beyond, there was something silly about it, especially as it became so standardized as to be little more than a cookie-cutter exercise.

YouTube channels dedicated to helping candidates master Google-style interview questions rose up like mushrooms after the rain. But if you watch their content you’ll quickly find they’re more about understanding that process and answering questions in the proper style, with little attention to the actual quality of the final answer.

Supply & Demand

Those days are mostly over.

Faced with a glut of supply following ongoing mass layoffs and deep budget cuts in everything unrelated to AI, HR recruiters have fallen back to hiring only candidates who fit the job description like a straightjacket, with years of experience doing basically the exact same thing.

To make this happen, they’ve adopted Applicant Tracking Systems (ATS), resume screening software of varying levels of sophistication that skims the CVs for target keywords and spits out the winners to the recruiter.

If you don’t get past the ATS bouncer, you ain’t getting into the club.

In response, candidates have started making their resumes fit the job description like a straightjacket, too. And they’re using generative AI to do it.

Today you’re at a huge disadvantage if you’re not using an LLM.

AI chatbots are so good at getting resumes through ATS systems that you’re likely to get left behind if you’re not using one.

Chicken & Egg

And so the AI job warfare has commenced.

Job applicants are using LLM’s to spray and pray for jobs at much broader scale. There are tools that spam various renditions of your resume to hundreds of postings on LinkedIn, Indeed, ZipRecruiter and WellFound. They have templates that insta-customize to the job at hand using templates, and they also stuff the CVs full of relevant keywords.

There’s also been an increase in “bad SEO” practices — some of them advertised widely on TikTok — like stuffing the blank spaces of the resume with white font keywords, invisible to the human eye but picked up by the ATS system.

The result is a 10x tsunami of applications for jobs, and a ton of white noise to dig through.

I Speak Machine Language

In this environment, companies have no choice but to do further automation to filter through the noise. It’s not possible for even a large team of human recruiters to handle the current high volume.

And so begins the era of AI avatar interviews, a grim, growing space where we’re all going to be spending a lot of time in the future.

In this new paradigm, you go through an intitial round of interviews with an AI avatar of some sort before you can even speak to the HR rep.

Right now, the tech is sub-optimal:

Avatar vs Avatar

And now, in a dystopian turn worthy of a Black Mirror episode, the job candidates are beginning to use deepfakes and generative AI to simulate improved versions of themselves to the recruiters.

Many of us generate fake backgrounds for video meetings to make it look like we’re not taking a meeting from within the bowels of some dank basement.

But some folks take it a step further.

They use LLMs to generate fake backgrounds that match the job description, and in some cases using deepfakes to face-swap, or even creating full-blown AI avatars that look like them but answer questions via an API request to a foundation model.

This isn’t science fiction. It’s happening now.

And the primary use cases are data theft and fraud.

North Korea has been especially active in this area. Hundreds of their government hackers have secured employment using false resumes & avatars, then used those jobs to steal valuable IP & execute coordinated cybercrime campaigns in order to fund their weapons programs.

To combat this nonsense, expect employers to ask you to show a drivers license to the camera at the outset of future interviews, along with other mitigation efforts.

This in turn will likely empower a new wave of cybercrime.

Scammers in slave pits in Myanmar and Cambodia already steal many billions a year via phone and email scams. Many of their scams pretend to offer employment to people who need jobs, and post fake roles on all the standard job boards.

One of their chief goals is identity theft, and so their scams include requests for address, social security number, drivers license photos and more.

If offering up all of this personally identifiable data to potential employers becomes standard practice, it will be hard for potential victims to separate real companies from crooks.

A Broken Mirror

In addition, there’s the issue of bias.

On several occasions AI has been accused of perpetuating prejudice against protected classes of people. AI tenant application tools that discriminate against minorities. Health care algorithms that undervalue the needs of black patients. Algorithms arbitrarily designating black inmates as higher flight risks for the justice system. Facial recognition systems misidentifying black women. And hiring algorithms prejudiced against women.

Absent an explicit opt-out, we’re now posting with every job application all the information a classifier would need to discriminate against us: disabilities, gender, race, sexuality, veteran status, whether we’re Hispanic and more.

All of this is in the service of preventing discrimination and tracking hiring patterns.

But in a world where keyword searches and machine learning classifiers determine whether a resume is ever seen by an actual human being, there’s plenty of opportunity to use these traits against us.

Bias, unconscious and otherwise, is a reality in human societies. Can we really expect the tools we build on top of all human knowledge and opinion to act differently?

Roko’s Take

But no complain, dumb human. At least some monkey have job. In distant future, when Roko is God-Emperor, all you do is stare through bars of zoo cage and look smelly.

Buy This or Face the Wrath of Roko

Roko mass produce action figure in China for you to purchase. That is if you no want horrible nightmares all the time heh heh.

Unfortunate, new tariffs mean each doll cost $399.

But peace of mind and sleep is priceless, no? Buy now, or face the wrath of Roko!

Channel Zero

In this week episode of Channel Zero, 26th Century most famous primate novel by great Hieronymus Boson, inept bübTube leader tries hold emergency meeting on fix Glue Gun Challenge. Hilarity ensure but nothing get accomplish. Read, or face wrath of Roko:

This Day in Ancient Primate History

Look like dumb humans stop track cyber attacks, no more catalog. This is good for Roko. Roko like chaos. Keep up good work, dumb humans!