Pivot to AI: Pay no attention to the man behind the curtain

By Amy Castor and David Gerard the September 12, 2023 for AmyCastor

“all this talk of AI xrisk has the stink of marketing too. Ronald McDonald telling people that he has a bunker in New Zealand because the new burger they’re developing in R&D might be so delicious society will crumble.”

— Chris Martin

Crypto’s being dull again — but thankfully, AI has been dull too. The shine is coming off. So we’re back on the AI beat.

The AI winter will be privatized

Since the buzzword “artificial intelligence” was coined in the 1950s, AI has gone through several boom and bust cycles.

A new technological approach looks interesting and gets a few results. It gets ridiculously hyped up and lands funding. The tech turns out to be not so great, so the funding gets cut. The down cycles are called AI winters.

Past AI booms were funded mainly by the US Department of Defense. But the current AI boom has been almost completely funded by venture capital.

The VCs who spent 2021 and 2022 pouring money into crypto startups are pivoting to AI startups, because people buy the idea that AI will change the world. In the first half of 2023, VCs invested more than $40 billion into AI startups, and $11 billion just in May 2023. This is even as overall VC funding for startups dropped by half in the same period from the year before. [Reuters; Washington Post]

The entire NASDAQ is being propped up by AI. It’s one of the only fields that is still hiring.

In contrast, the DOD only requested $1.8 billion for AI funding in its 2024 budget. [DefenseScoop]

So why are VCs pouring money into AI?

Venture capital is professional gambling. VCs are looking for a liquidity event. One big winner can pay for a lot of failures.

Finding someone to buy a startup you’ve funded takes marketing and hype. The company doing anything useful, or anything that even works, is optional.

What’s the exit plan for AI VCs? Where’s the liquidity event? Do they just hope the startups they fund will do an initial public offering or just get acquired by a tech giant before the market realizes AI is running out of steam?

We’re largely talking about startups whose business model is sending queries to OpenAI.

At least with “Web3,” the VCs would just dump altcoins on retail investors via their very good friends at Coinbase. But with AI, we can’t see an obvious exit strategy beyond finding a greater fool.

Pay no attention to the man behind the curtain

The magical claim of machine learning is that if you give the computer data, the computer will work out the relations in the data all by itself. Amazing!

In practice, everything in machine learning is incredibly hand-tweaked. Before AI can find patterns in data, all that data has to be tagged, and output that might embarrass the company needs to be filtered.

Commercial AI runs on underpaid workers in English-speaking countries in Africa creating new training data and better responses to queries. It’s a painstaking and laborious process that doesn’t get talked about nearly enough.

The workers do individual disconnected actions all day, every day — so called “tasks” — working for companies like Remotasks, a subsidiary of Scale AI, and doing a huge amount of the work behind OpenAI.

AI doesn’t remove human effort. It just makes it much more alienated.

There’s an obvious hack here. If you are an AI task worker, your goal is to get paid as much as possible without too much effort. So why not use some of the well-known tools for this sort of job? [New York]

Another Kenyan annotator said that after his account got suspended for mysterious reasons, he decided to stop playing by the rules. Now, he runs multiple accounts in multiple countries, tasking wherever the pay is best. He works fast and gets high marks for quality, he said, thanks to ChatGPT. The bot is wonderful, he said, letting him speed through $10 tasks in a matter of minutes. When we spoke, he was having it rate another chatbot’s responses according to seven different criteria, one AI training the other.

Remember, the important AI use case is getting venture capital funding. Why buy or rent expensive computing when you can just pay people in poor countries to fake it? Many “AI” systems are just a fancier version of the original Mechanical Turk.

Facebook’s M from 2017 was an imitation of Apple’s Siri virtual assistant. The trick was that hard queries would be punted to a human. Over 70% of queries ended up being answered by a human pretending to be the bot. M was shut down a year after launch.

Kaedim is a startup that claims to turn two-dimensional sketches into 3-D models using “machine learning.” The work is actually done entirely by human modelers getting paid $1-$4 per 15-minute job. But then, the founder, Konstantina Psoma, was a Forbes 30 Under 30. [404 Media; Forbes]

The LLM is for spam

OpenAI’s AI-powered text generators fueled a lot of the hype around AI — but the real-world use case for large language models is overwhelmingly to generate content for spamming. [Vox]

The use case for AI is spam web pages filled with ads. Google considers LLM-based ad landing pages to be spam, but seems unable or unwilling to detect and penalize it. [MIT Technology Review; The Verge]

The use case for AI is spam books on Amazon Kindle. Most are “free” Kindle Unlimited titles earning money through subscriber pageviews rather than outright purchases. [Daily Dot]

The use case for AI is spam news sites for ad revenue. [NewsGuard]

The use case for AI is spam phone calls for automated scamming — using AI to clone people’s voices. [CBS]

The use case for AI is spam Amazon reviews and spam tweets. [Vice]

The use case for AI is spam videos that advertise malware. [DigitalTrends]

The use case for AI is spam sales sites on Etsy. [The Atlantic, archive]

The use case for AI is spam science fiction story submissions. Clarkesworld had to close submissions because of the flood of unusable generated garbage. The robot apocalypse in action. [The Register]

Supertoys last all summer long

End users don’t actually want AI-based products. Machine learning systems can generate funny text and pictures to show your friends on social media. But even that’s wearing thin — users mostly see LLM output in the form of spam.

LLM writing style and image generator drawing style are now seen as signs of low quality work. You can certainly achieve artistic quality with AI manipulation, as in this music video — but even this just works on its novelty value. [YouTube]

For commercial purposes, the only use case for AI is still to replace quality work with cheap ersatz bot output — in the hope of beating down labor costs.

Even then, the AI just isn’t up to the task.

Microsoft put $10 billion into OpenAI. The Bing search engine added AI chat — and it had almost no effect on user numbers. It turns out that search engine users don’t want weird bot responses full of errors. [ZDNet]

The ChatGPT website’s visitor numbers went down 10% in June 2023. LLM text generators don’t deliver commercial results, and novelty only goes so far. [Washington Post]

After GPT-3 came out, OpenAI took three years to make an updated version. GPT-3.5 was released as a stop-gap in October 2022. Then GPT-4 finally came out in March 2023! But GPT-4 turns out to be eight instances of GPT-3 in a trenchcoat. The technology is running out of steam. [blog post; Twitter, archive]

Working at all will be in the next version

The deeper problem is that many AI systems simply don’t work. The 2022 paper “The fallacy of AI functionality” notes that AI systems are often “constructed haphazardly, deployed indiscriminately, and promoted deceptively.”

Still, machine learning systems do some interesting things, a few of which are even genuinely useful. We asked GitHub and they told us that they encourage their own employees to use the GitHub Copilot AI-based autocomplete system for their own internal coding — with due care and attention. We know of other coders who find Copilot to be far less work than doing the boilerplate by hand.

(Though Google has forbidden its coders from using its AI chatbot, Bard, to generate internal code.) [The Register]

Policy-makers and scholars — not just the media — tend to propagate AI hype. Even if they try to be cautious, they may work in terms of ethics of deployment, and presume that the systems do what they’re claimed to do — when they often just don’t.

Ethical considerations come after you’ve checked basic functionality. Always put functionality first. Does the system work? Way too often, it just doesn’t. Test and measure. [arXiv, PDF, 2022]

AI is the new crypto mining

In 2017, the hot buzzword was “blockchain” — because the price of bitcoin was going up. Struggling businesses would add the word “blockchain” to their name or their mission statement, in the hope their stock price would go up. Long Island Iced Tea became Long Blockchain and saw its shares surge 394%. Shares in biotech company Bioptix doubled in price when it changed its name to Riot Blockchain and pivoted to bitcoin mining. [Bloomberg, 2017, archive; Bloomberg, 2017, archive]

The same is now happening with AI. Only it’s not just the venture capitalists — even the crypto miners are pivoting to AI.

Bitcoin crashed last year and crypto mining is screwed. As far as we can work out, the only business plan was to get foolish investors’ money during the bubble, then go bankrupt.

In mid-2024, the bitcoin mining reward will halve again. So the mining companies are desperate to find other sources of income.

Ethereum moved to proof of stake in September 2022 and told its miners to just bugger off. Ethereum was mined on general-purpose video cards — so miners have a glut of slightly-charred number crunching machinery.

Hive Blockchain in Vancouver is pivoting to AI to repurpose its pile of video cards. It’s also changed its name to Hive Digital Technologies. [Bloomberg, archive; press release]

Marathon Digital claims that “over time you’re going to see that blockchain technologies and AI have a very tight coupling.” No, us neither. Marathon is doubling and tripling down on bitcoin mining — but, buzzwords! [Decrypt]

Nvidia makes the highest-performance video cards. The GPU processors on these cards turn out to be useful for massively parallel computations in general — such as running the calculations needed to train machine learning models. Nvidia is having an excellent year and its market cap is over $1 trillion.

So AI can take over from crypto in yet another way — carbon emissions from running all those video cards.

AI’s massive compute load doesn’t just generate carbon — it uses huge amounts of fresh water for cooling. Microsoft’s water usage went up 34% between 2021 and 2022, and they blame AI computation. ChatGPT uses about 500 mL of water every time you have a conversation with it. [AP]

We don’t yet have a Digiconomist of AI carbon emissions. Go start one.

Cybersecurity is broken

27 March 2024

It is a well-known fact that we dish out a whole lot of shit talk around these parts. And by "we" I mean me, but that's beside the point. Talking smack about 10-ply LinkedIn vCISOs is, quite honestly, pretty easy and kind of satisfying because some 8 out of 10 times they are stupid as fuck and deserve the heckling. The remaining 2 out of 10 are maybe trying to fight the good fight, and do right by their teams. Maybe. Don't you quote me on that figure. Actually, best you don't quote me at all because there are peeps out there saying things that are much more clever. Take this quote(?) from one Bob Metcalfe (tks, snowcrasher!)

"The Stockings Were Hung by the Chimney with Care"

The ARPA Computer Network is susceptible to security violations for at least
the three following reasons:

(1) Individual sites, used to physical limitations on machine access, have
not yet taken sufficient precautions toward securing their systems
against unauthorized remote use. For example, many people still use
passwords which are easy to guess: their fist names, their initials,
their host name spelled backwards, a string of characters which are
easy to type in sequence (e.g. ZXCVBNM).

(2) The TIP allows access to the ARPANET to a much wider audience than
is thought or intended. TIP phone numbers are posted, like those
scribbled hastily on the walls of phone booths and men's rooms. The
TIP required no user identification before giving service. Thus,
many people, including those who used to spend their time ripping off
Ma Bell, get access to our stockings in a most anonymous way.

(3) There is lingering affection for the challenge of breaking
someone's system. This affection lingers despite the fact that
everyone knows that it's easy to break systems, even easier to
crash them.

All of this would be quite humorous and cause for raucous eye
winking and elbow nudging, if it weren't for the fact that in
recent weeks at least two major serving hosts were crashed
under suspicious circumstances by people who knew what they
were risking; on yet a third system, the system wheel password
was compromised -- by two high school students in Los Angeles
no less.

We suspect that the number of dangerous security violations is
larger than any of us know is growing. You are advised
not to sit "in hope that Saint Nicholas would soon be there".

That's from 1973. The dude who invented Ethernet was worried about what we now call cybersecurity fifty fucking years ago. Several wake-up calls happened since then: phreaking peeps exploding the phones, hacker supergroups testifying in front of the US Senate on the topic of cybersecurity, hacker supergroups releasing super easy to use RATs, a cornucopia of malware, including shit made by nation-states, and ransomware attacks that are only profitable because some people just decided that an inefficient distributed database was worth some money. A lot of those issues were only made possible by people's insistence on using a programming language from half a century ago when better options are available. And that's just the technical side of things.

Take, for example, the Pen Test Partners' research on Peloton's API security. The researchers were able to grab a whole bunch of data that was supposed to be private, disclosed the issue to Peloton who, in turn, basically ghosted the researcher until a TechCrunch reporter got involved. Classic case of "we're not sorry we suck at protecting our customers' data, we're just sorry we got caught." I mean, if you need to get fucking TechCrunch involved to be taken seriously, the situation is hopeless.

Absolutely no amount of gentle pleas disguised as executive orders from the White House urging people to use memory-safe languages will solve the problem. CISA, despite all the phenomenal work they do, can't charge people who mishandle data with negligence; critical infrastructure involved or not. And maybe they should.

You see, cybersecurity is broken because of the lack of consequences. It's really that simple. When literally nothing happens when some stupid service gets popped and loses your data they had no business collecting in the first place, this kind of thing will happen over and over and over again. Why the fuck do you need my home address just so I can copy and paste some GIFs? Because you want to sell this data to data brokers, and you know there will be absolutely no negative consequences if you mishandle this data, fucking over the people who keep your business afloat. So, companies big and small fuck things up and we need to clean up the mess and face the consequences. Sounds about right.

Cybersecurity is even more broken when these companies that face zero consequences look at their payroll and think "Wait a fucking minute! Why the hell are we spending six full dollars a year on cybersecurity when we can, I dunno, do nothing at all for free because cybersecurity incidents will not negatively impact our bottomline whatsoever?" That's why you, my cybersecurity and infosec brethren, are getting laid off. That's why you don't have the tools you need. That's why you don't get the training you should. That's why you're overworked. That's why you're stuck as an underpaid individual contributor doing the work of 5 people for $75k/year while your CISO who makes $500k is on LinkedIn all day writing stupid shit about AI.

Cybersecurity is broken because fixing it benefits no one but the regular, unremarkable, salt of the earth folks. And, according to the powers that be, fuck them folks. Fixing it requires strong data protection laws, but passing laws is just not something the overwhelming majority of legislative bodies in the world do. Passing laws that slightly inconvenience multi-billion dollar corporations while improving the lives of the plebes is even more of a tall order. And that's true for a whole lot of things that have nothing to do with cybersecurity, but this is a blog about cybersecurity, so please bear with me.

That's the answer: data protection laws. You get my data for rEaSoNs, and you fuck it up? You should pay a hefty price that cannot be written off as the cost of doing business. You make data brokers illegal, or, at the very least, way less profitable. You do what the payment card industry has been doing for decades: you tell everyone handling your data that they have to follow a very comprehensive set of data protection rules, lest they get fined or cut off entirely. A group of four credit card companies can do that, so I'm sure mighty governments can, too.

But how do we push things in the right direction? Well, that's one of the many topics we discuss in our Discord server (Hey you guys!). Not only are my fellow Crankies inspiring the shit out of me every day, we have bigger plans than just shitposting and commiserating. Turns out that buying a congressperson lobbying is not that expensive, really. We are working on something that we hope will help lift everyone in this industry up. As I once wrote on that very Discord: "When we abstain from using our collective power of influence, we lose by default." Or "you miss 100% of the shots you don't take" or whatever the fuck Gretzky said. We're about 700-strong and planning on doing great things. Come join us because the movement cannot be stopped.

Previous: Pigeons As Far As The Eye Can See

Twitter is becoming a 'ghost town' of bots as AI-generated spam content floods the internet

ABC Science / By technology reporter James Purtill

Parts of the web are now dominated by bots and junk websites designed to go unread by humans.

One morning in January this year, marine scientist Terry Hughes opened X (formerly Twitter) and searched for tweets about the Great Barrier Reef.

"I keep an eye on what's being tweeted about the reef every day," Professor Hughes, a leading coral researcher at James Cook University, said.

What he found that day surprised and confused him; hundreds of bot accounts tweeting the same strange message with slightly different wording.

"Wow, I had no idea that agricultural runoff could have such a devastating impact on the Great Barrier Reef," one account, which otherwise spruiked cryptocurrencies, tweeted.

Another crypto bot wrote: "Wow, it's disheartening to hear about the water pollution challenges Australia faces."

And so on. Hundreds of crypto accounts tweeting about agricultural runoff.

A month later, it happened again. This time, bots were tweeting about "marine debris" threatening the Great Barrier Reef.

What was going on?

When Professor Hughes tweeted what he'd found, some saw a disinformation conspiracy, an attempt to deflect attention from climate change.

The likely answer, however, is more mundane, but also more far-reaching in its implications.

More than a year since Elon Musk bought X with promises to get rid of the bots, the problem is worse than ever, experts say.

And this is one example of a broader problem affecting online spaces.

The internet is filling up with "zombie content" designed to game algorithms and scam humans.

It's becoming a place where bots talk to bots, and search engines crawl a lonely expanse of pages written by artificial intelligence (AI).

Junk websites clog up Google search results. Amazon is awash with nonsense e-books. YouTube has a spam problem.

And this is just a trickle in advance of what's been called the "great AI flood".

Bots liking bots, talking to other bots

But first, let's get back to those reef-tweetin' bots.

Timothy Graham, an expert on X bot networks at the Queensland University of Technology, ran the tweets through a series of bot and AI detectors.

Dr Graham found 100 per cent of the text was AI-generated.

"Overall, it appears to be a crypto bot network using AI to generate its content," he said.

"I suspect that at this stage it's just trying to recruit followers and write content that will age the fake accounts long enough to sell them or use them for another purpose."

That is, the bots probably weren't being directed to tweet about the reef in order to sway public opinion.

Dr Graham suspects these particular bots probably have no human oversight, but are carrying out automated routines intended to out-fox the bot-detection algorithms.

Searching for meaning in their babble was often pointless, he said.

"[Professor Hughes] is trying to interpret it and is quite right to try and make sense of it, but it just chews up attention, and the more engagement they get, the more they are rewarded.

The cacophony of bot-talk degrades the quality of online conversations. They interrupt the humans and waste their time.

"Here's someone who is the foremost research scientist in this space, spending their time trying to work out the modus operandi of these accounts."

In this case, the bots were replying to the tweet of another bot, which, in turn, replied to the tweets of other bots, and so on.

One fake bot account was stacked on top of the other, Dr Graham said.

"It's AI bots all the way down."

How bad is X's bot problem?

In January, a ChatGPT glitch appeared to shine a light on X's bot problem.

For a brief time, some X accounts posted ChatGPT's generic response to requests that it deems outside of its content policy, exposing them as bots that use ChatGPT to generate content.

Users posted videos showing scrolling feeds with numerous accounts stating "I'm sorry, but I cannot provide a response to your request as it goes against OpenAl's content policy."

"Twitter is a ghost town," one user wrote.

But the true scale of X's bot problem is difficult for outsiders to estimate.

Shortly after Mr Musk gained control of X while complaining about bots, X shut down free access to the programming interface that allowed researchers to study this problem.

That left researchers with two options: pay X for access to its data or find another way to peek inside.

Towards the end of last year, Dr Graham and his colleagues at QUT paid X $7,800 from a grant fund to analyse 1 million tweets surrounding the first Republican primary debate.

They found the bot problem was worse than ever, Dr Graham said at the time.

Later studies support this conclusion. Over three days in February, cybersecurity firm CHEQ tracked the proportion of bot traffic from X to its clients' websites.

It found three-quarters of traffic from X was fake, compared to less than 3 per cent of traffic from each of TikTok, Facebook and Instagram.

"Terry Hughes' experience is an example of what's going on on the platform," Dr Graham said.

"One in 10 likes are from a porn bot, anecdotally."

The rise of a bot-making industry

So what's the point of all these bots? What are they doing?

Crypto bots drive up demand for certain coins, porn bots get users to pay for porn websites, disinformation bots peddle fake news, astroturfing bots give the impression of public support, and so on.

Some bots exist purely to increase the follower counts and engagement statistics of paying customers.

A sign of the scale of X's bot problem is the thriving industry in bot-making.

Bot makers from around the world advertise their services on freelancer websites.

Awais Yousaf, a computer scientist in Pakistan, sells "ChatGPT Twitter bots" for $30 to $500, depending on their complexity.

In an interview with the ABC, the 27-year-old from Gujranwala said he could make a "fully fledged" bot that could "like comments on your behalf, make comments, reply to DMs, or even make engaging content according to your specification".

Mr Yousaf's career tracks the rise of the bot-making economy and successive cycles of internet hype.

Having graduated from university five years ago, he joined Pakistan's growing community of IT freelancers from "very poor backgrounds".

Many of the first customers wanted bots to promote cryptocurrencies, which were booming in popularity at the time.

"Then came the NFT thing," he said.

A few years ago he heard about OpenAI's GPT3 language model and took a three-month break to learn about AI.

"Now, almost 90 per cent of the bots I do currently are related to AI in one way or another.

"It can be as simple as people posting AI posts regarding fitness, regarding motivational ideas, or even cryptocurrency predictions."

In five years he's made 120 Twitter bots.

Asked about Mr Musk's promise to "defeat the spam bots," Mr Yousaf smiled.

"It's hard to remove Twitter bots from Twitter because Twitter is mostly bot."

AI-generated spam sites may overwhelm search engines

X's bot problem may be worse than other major platforms, but it's not alone.

A growing "deluge" of AI content is flooding platforms that were "never designed for a world where machines can talk with people convincingly", Dr Graham said.

"It's like you're running a farm and had never heard of a wolf before and then suddenly you have new predators on the scene.

"The platforms have no infrastructure in place. The gates are open."

The past few months have seen several examples of this.

Companies are using AI to rewrite other media outlet's stories, including the ABC's, to then publish them on the company's competing news websites.

A company called Byword claims it stole 3.6 million in "total traffic" from a competitor by copying their site and rewriting 1,800 articles using AI.

"Obituary pirates" are using AI to create YouTube videos of people summarising the obituaries of strangers, sometimes fabricating details about their deaths, in order to capture search traffic.

Authors are reporting what appear to be AI-generated imitations and summaries of their books on Amazon.

Google's search results are getting worse due to spam sites, according to a recent pre-print study by German researchers.

The researchers studies search results for thousands of product-review terms across Google, Bing and DuckDuckGo over the course of a year.

They found that higher-ranked pages tended to have lower text quality but were better designed to game the search ranking algorithm.

"Search engines seem to lose the cat-and-mouse game that is SEO spam," they wrote in the study.

Co-author Matti Wiegman from Bauhaus University, Weimar said this rankings war was likely to get much worse with the advent of AI-generated spam.

"What was previously low-quality content is now very difficult to distinguish from high-quality content," he said.

"As a result, it might become difficult to distinguish between authentic and trustworthy content that is useful and content that is not."

He added that the long-term effects of AI-generated content on search engines was difficult to judge.

AI-generated content could make search more useful, he said.

"One possible direction is that generated content will become better than the low-quality human-made content that dominates some genres in web search, in which case the search utility will increase."

Or the opposite will happen. AI-generated content will overwhelm "vulnerable spaces" such as search engines and "broadcasting-style" social media platforms like X.

In their place, people may turn to "walled gardens" and specialised forums with smaller numbers of human-only members.

Platforms prepare for coming flood

In response to this emerging problem, platforms are trialling different strategies.

Meta recently announced it was building tools to detect and label AI-generated images posted on its Facebook, Instagram and Threads services.

Amazon has limited authors to uploading a maximum of three books to its store each day, although authors say that hasn't solved the problem.

X is trialling a "Not a Bot" program in some countries where it charges new users $1 per year for basic features.

This program operates alongside X's verification system, where users pay $8 per month to have their identity checked and receive a blue tick.

But it appears the bot-makers have found a way around this.

All the reef-tweeting crypto bots Professor Hughes found were verified accounts.

"It's clutter on the platform that's not necessary. You'd wish they'd clean it up," the coral scientist said.

"It wastes everyone's time."

Losing the imitation game

AI cannot develop software for you, but that's not going to stop people from trying to make it happen anyway. And that is going to turn all of the easy software development problems into hard problems.

If you've been anywhere near major news or social media in the last few months, you've probably heard repeatedly about so-called AI, ChatGPT, and large language models (LLMs). The hype surrounding these topics has been intense. And the rhetoric has been manipulative, to say the least. Proponents have claimed that their models are or soon will be generally intelligent, in the way we mean humans are intelligent. They're not. They've claimed that their AI will eliminate whole categories of jobs. And they've claimed that developing these systems further and faster is both necessary and urgent, justified by science fiction dressed up as arguments for some sort of "safety" that I find to be incoherent.

The outer layer of hype surrounding AI—and LLM chatbots in particular—is that they will become indispensable tools of daily work, and entirely replace people in numerous categories of jobs. These claims have included the fields of medicine, law, and education, among others. I think it's nonsense. They imagine self-teaching classrooms and self-diagnosing fitness gadgets. These things will probably not even work as well as self-driving cars, which is to say: only well enough to be dangerous. Of course, that's not stopping people from pushing these fantasies, anyway. But these fields are not my area of expertise. My expertise is in software engineering. We should know better, but software developers are falling victim to the same kind of AI fantasies.

A computer can never be held accountable. Therefore, a computer must never make a management decision.

While the capabilities are fantasy, the dangers are real. These tools have denied people jobs, housing, and welfare. All erroneously. They have denied people bail and parole, in such a racist way it would be comical if it wasn't real. And the actual function of AI in all of these situations is to obscure liability for the harm these decisions cause.

So-Called AI

Artificial Intelligence is an unhelpful term. It serves as a vehicle for people's invalid assumptions. It hand-waves an enormous amount of complexity regarding what "intelligence" even is or means. And it encourages people confuse concepts like cognition, agency, autonomy, sentience, consciousness, and a host of related ideas. However, AI is the vernacular term for this whole concept, so it's the one I'll use. I'll let other people push that boulder, I'm here to push a different one.

Those concepts are not simple ideas, either. Describing them gets into hard questions of psychology, neurology, anthropology, and philosophy. At least. Given that these are domains that the tech field has routinely dismissed as unimportant for decades, maybe it shouldn't be surprising that techies as a group are now completely unprepared to take a critical view of claims about AI.

The Turing Test

Certainly part of how we got here is the Turing test. That is, the pop science reduction of Alan Turing's imitation game. The actual proposal is more substantial. And taking it seriously produces some interesting reading. But the common notion is something like a computer is intelligent if it can reliably pass as human in conversation. I hope seeing it spelled out like that makes it clear how dramatically that overreaches. Still, it's the framework that people have, and it informs our situation. I think the bit that is particularly informative is the focus on natural, conversational language. And also, the deception inherent in the imitation game scenario, but I'll come back to that.

Our understanding of intelligence is a moving target. We only have one meaningful fixed point to work from. We assert that humans are intelligent. Whether anything else is, is not certain. What intelligence itself is, is not certain. Not too long ago, a lot of theory rested on our ability to create and use tools. But then that ability turned out to be not as rare as we thought, and the consensus about the boundaries of intelligence shifted. Lately, it has fallen to our use of abstract language. That brings us back to AI chatbots. We suddenly find ourselves confronted with machines that seem to have a command of the English language that rivals our own. This is unfamiliar territory, and at some level it's reasonable that people will reach for explanations and come up with pop science notions like the Turing test.

Language: any system of formalized symbols, signs, sounds, gestures, or the like used or conceived as a means of communicating thought, emotion, etc.

Language Models

ChatGPT and the like are powered by large language models. Linguistics is certainly an interesting field, and we can learn a lot about ourselves and each other by studying it. But language itself is probably less than you think it is. Language is not comprehension, for example. It's not feeling, or intent, or awareness. It's just a system for communication. Our common lived experiences give us lots of examples that anything which can respond to and produce common language in a sensible-enough way must be intelligent. But that's because only other people have ever been able to do that before. It's actually an incredible leap to assume, based on nothing else, that a machine which does the same thing is also intelligent. It's much more reasonable to question whether the link we assume exists between language and intelligence actually exists. Certainly, we should wonder if the two are as tightly coupled as we thought.

That coupling seems even more improbable when you consider what a language model does, and—more importantly—doesn't consist of. A language model is a statistical model of probability relationships between linguistic tokens. It's not quite this simple, but those tokens can be thought of as words. They might also be multi-word constructs, like names or idioms. You might find "raining cats and dogs" in a large language model, for instance. But you also might not. The model might reproduce that idiom based on probability factors instead. The relationships between these tokens span a large number of parameters. In fact, that's much of what's being referenced when we call a model large. Those parameters represent grammar rules, stylistic patterns, and literally millions of other things.

What those parameters don't represent is anything like knowledge or understanding. That's just not what LLMs do. The model doesn't know what those tokens mean. I want to say it only knows how they're used, but even that is over stating the case, because it doesn't know things. It models how those tokens are used. When the model works on a token like "Jennifer", there are parameters and classifications that capture what we would recognize as things like the fact that it's a name, it has a degree of formality, it's feminine coded, it's common, and so on. But the model doesn't know, or understand, or comprehend anything about that data any more than a spreadsheet containing the same information would understand it.

Mental Models

So, a language model can reproduce patterns of language. And there's no particular reason it would need to be constrained to natural, conversational language, either. Anything that's included in the set of training data is fair game. And it turns out that there's been a lot of digital ink spent on writing software and talking about writing software. Which means those linguistic patterns and relationships can be captured and modeled just like any other. And sure, there are some programming tasks where just a probabilistic assembly of linguistic tokens will produce a result you want. If you prompt ChatGPT to write a python function that fetches a file from S3 and records something about it in DynamoDB, I would bet that it just does, and that the result basically works. But then, if you prompt ChatGPT to write an authorization rule for a new role in your application's proprietary RBAC system, I bet that it again just does, and that the result is useless, or worse.

Programming as Theory Building

Non-trivial software changes over time. The requirements evolve, flaws need to be corrected, the world itself changes and violates assumptions we made in the past, or it just takes longer than one working session to finish. And all the while, that software is running in the real world. All of the design choices taken and not taken throughout development; all of the tradeoffs; all of the assumptions; all of the expected and unexpected situations the software encounters form a hugely complex system that includes both the software itself and the people building it. And that system is continuously changing.

The fundamental task of software development is not writing out the syntax that will execute a program. The task is to build a mental model of that complex system, make sense of it, and manage it over time.

To circle back to AI like ChatGPT, recall what it actually does and doesn't do. It doesn't know things. It doesn't learn, or understand, or reason about things. What it does is probabilistically generate text in response to a prompt. That can work well enough if the context you need to describe the goal is so simple that you can write it down and include it with the prompt. But that's a very small class of essentially trivial problems. What's worse is there's no clear boundary between software development problems that are trivial enough for an LLM to be helpful vs being unhelpful. The LLM doesn't know the difference, either. In fact, the LLM doesn't know the difference between being tasked to write javascript or a haiku, beyond the different parameters each prompt would activate. And it will readily do a bad job of responding to either prompt, with no notion that there even is such a thing as a good or bad response.

Software development is complex, for any non-trivial project. But complexity is hard. Overwhelmingly, when we in the software field talk about developing software, we've dealt with that complexity by ignoring it. We write code samples that fit in a tweet. We reduce interviews to trivia challenges about algorithmic minutia. When we're feeling really ambitious, we break out the todo app. These are contrivances that we make to collapse technical discussions into an amount of context that we can share in the few minutes we have available. But there seem to be a lot of people who either don't understand that or choose to ignore it. They frame the entire process of software development as being equivalent to writing the toy problems and code samples we use among general audiences.

Automating the Easy Part

The intersection of AI hype with that elision of complexity seems to have produced a kind of AI booster fanboy, and they're making personal brands out of convincing people to use AI to automate programming. This is an incredibly bad idea. The hard part of programming is building and maintaining a useful mental model of a complex system. The easy part is writing code. They're positioning this tool as a universal solution, but it's only capable of doing the easy part. And even then, it's not able to do that part reliably. Human engineers will still have to evaluate and review the code that an AI writes. But they'll now have to do it without the benefit of having anyone who understands it. No one can explain it. No one can explain what they were thinking when they wrote it. No one can explain what they expect it to do. Every choice made in writing software is a choice not to do things in a different way. And there will be no one who can explain why they made this choice, and not those others. In part because it wasn't even a decision that was made. It was a probability that was realized.

[A programmer's] education has to emphasize the exercise of theory building, side by side with the acquisition of knowledge of data processing and notations.

But it's worse than AI being merely inadequate for software development. Developing that mental model requires learning about the system. We do that by exploring it. We have to interact with it. We manipulate and change the system, then observe how it responds. We do that by performing the easy, simple programing tasks. Delegating that learning work to machines is the tech equivalent of eating our seed corn. That holds true beyond the scope of any team, or project, or even company. Building those mental models is itself a skill that has to be learned. We do that by doing it, there's not another way. As people, and as a profession, we need the early career jobs so that we can learn how to do the later career ones. Giving those learning opportunities to computers instead of people is profoundly myopic.

Imitation Game

If this is the first time you're hearing or reading these sentiments, that's not too surprising. The marketing hype surrounding AI in recent months has been intense, pervasive, and deceptive. AI is usually cast as being hyper competent, and superhuman. To hear the capitalists who are developing it, AI is powerful, mysterious, dangerous, and inevitable. In reality, it's almost none of those things. I'll grant that AI can be dangerous, but not for the reasons they claim. AI is complicated and misunderstood, and this is by design. They cloak it in rhetoric that's reminiscent of the development of atomic weapons, and they literally treat the research like an arms race.

I'm sure there are many reasons they do this. But one of the effects it has is to obscure the very mundane, serious, and real harms that their AI models are currently perpetuating. Moderating the output of these models depends on armies of low paid and precariously employed human reviewers, mostly in Kenya. They're subjected to the raw, unfiltered linguistic sewage that is the result of training a language model on uncurated text found on the public internet. If ChatGPT doesn't wantonly repeat the very worst of the things you can find on reddit, 4chan, or kiwi farms, that is because it's being dumped on Kenyan gig workers instead.

That's all to say nothing of the violations of intellectual property and basic consent that was required to train the models in the first place. The scale of the theft and exploitation required to build the data sets these models train with is almost inconceivable. And the energy consumption and e-waste produced by these systems is staggering.

All of this is done to automate the creation of writing or media that is designed to deceive people. It's intended to seem like people, or like work done by people. The deception, from both the creators and the AI models themselves, is pervasive. There may be real, productive uses for these kinds of tools. There may be ways to build and deploy them ethically and sustainably. But that's not the situation with the instances we have. AI, as it's been built today, is a tool to sell out our collective futures in order to enrich already wealthy people. They like to frame it as being akin to nuclear science. But we should really see it as being more like fossil fuels