Bobby Hiltz

NOTE: Some of the articles mentioned in this post were selected at random from my RSS feeds between the 23rd and 24th of April 2026. I didn’t take time to cherry-pick anything. Other articles came from my long list of bookmarked articles on Wallabag.

Motivation

The news is awash with headlines making claims about AI.

But, how do I — a standard issue, non-dev, mouth-breathing, normie human — understand all of this?

This is a paraphrase of a common plight among my students (and some colleagues (and myself)). “The news said this,” they say, “how am I supposed to understand or digest this?”

“With handfuls of salt!” I often tell them (and myself).

The real answer is to use your noggin (It’s the original thought computer). It would be nice if I could pat myself on the back and say, “good job dude, your critical thinking skills are through the roof!” But, that would be a lie. The truth is that I spend too much time online and a lot of that is reading the news. The news is dominated by very American-centric stories, war, genocide, entertainment gossip, and “Artificial Intelligence.”

It would be impossible (and illogical, if we’re being honest) for me to take the time to research and study the methods of reporting on AI. Beginning with how a journalist could understand AI and report on it if scientists don’t know how it works? They don’t, I suppose. Depending on how you look at it, that could be perceived negatively. For me, though, it leaves the door open for a “two can play” scenario.

If a journalist, who doesn’t know anything about AI, can write about AI, then I, a person who doesn’t know a thing about journalism, can write about journalism!

ANOTHER NOTE: Many journalists certainly do understand more than what is included in their published articles. Could you imagine if every article about AI included a paragraph or more about how AI isn’t living up to the hype or how the CEO is not to be trusted? It would be like if every article about Elon Musk reminded us that he has had hair implants or that he is a fascist. They should be reminding us, but they would lose readers.

The way of the words

Newspapers and websites exist to be engaged with. Online, that means clicks and other KPIs and jargon from the world of online advertising. Headlines are designed to be attractive, or scary, or unbelievable. Table your emotions, because your first reaction should be one of doubt and suspicion. This is referred to as FUD in many online circles: Fear, Uncertainty, and Doubt. A half-assed listicle titled “5 jobs that AI can already do better” will have you worrying about your future.

Also, companies like OpenAI and Anthropic have been trying to turn their chatbots into products since 2022. Their investors and early adopters will always promote and anthropomorphize. If it sounds too good to be true, too revolutionary, too insane, then it is just speculation. “Is it time to start considering personhood rights for AI chatbots?” was the title of an opinion piece in the LA Times in March 2023.

Headline grammar is straightforward. Is it past, present, or future tense? Did it supposedly just happen, is it happening now or is it some sort of maybe soon situation? This is important because it hints at what emotion the article is trying to evoke. The word choice, too. Here is an example:

• AI failure could trigger the next financial crisis, warns Elizabeth Warren (The Verge)

“Could” is less certain than “can,” and the words “failure,” “trigger,” “crisis,” and “warns” all have negative connotations. This is a sound bite turned into a story. If you read the story (five short paragraphs) you’ll find that the full quote isn’t really news; it is a rephrasing of what even proponents of AI will admit:

“If AI companies are unable to increase revenues with lightning speed, they won’t be able to service their massive debt loads, [a]nd because of shady accounting strategies, the first big stumble will have everyone running for the exits, potentially triggering destabilizing losses in the financial sector and another 2008-style financial crisis.”

If dissecting grammar isn’t your bag, treat it like bingo; put the articles in boxes, or categories. You’ll need to make your own categories, but brains are pretty good at sorting things into lists and seeing patterns. I don’t have a fixed set of categories, over time the headlines have become a never-ending collection of overlapping tropes.

Today’s tropes

Newspapers are fans of telling us that AI will replace humans (look for words like “agent”) or that AI cannot replace certain humans:

• OpenAI now lets teams make custom bots that can do work on their own (The Verge)
• Sean Duffy denies replacing air traffic controllers with AI (The Hill)
• ChatGPT 5.5 Is All About Math, Science and AI Research (CNET)

Other popular tropes in this category I’ve noticed are “training your own replacement” and “AI-proof careers.”

You need to use AI because criminals use AI and if you don’t learn now you’ll fall behind (look for words like “risk” and “danger” and “cybersecurity”):

• Anthropic’s most dangerous AI model just fell into the wrong hands (The Verge)
• House lawmakers get a chilling demo of ‘jailbroken’ AI (POLITICO)
• AI is 10 to 20 times more likely to help you build a bomb if you hide your request in cyberpunk fiction, new research paper says (PC Gamer)
• The Mythos meeting focused on the wrong AI risk to banks. Here’s the one nobody is talking about (Fortune)

Follow me off this cliff ([BigCompany] is doing something with AI):

• Visa CMO: AI agents are your new customers — here’s how to sell to them Fortune)
• Yelp’s AI chatbot can now make your dinner reservation (Endgadget)

Holy frakking shit, it’s bunker time baby!:

• AI failure could trigger the next financial crisis, warns Elizabeth Warren (The Verge)

An AI-bro or CEO says (see Karl Bode’s “CEO Said A Thing!” Journalism):

• iPhone grows in importance as AI improves, Perplexity CEO says (Apple Insider)
• ‘An AI-led defense strategy that’s overseen by humans’: Google is introducing more agents to its ‘full AI stack’ to allow AI security at ‘infinite scale’ (TechRadar)

An issue with making lists is that the number of lists will eventually become uncountable. You’ll soon need to make sublists. As good at sorting things as the brain is, it is also lazy and there is a simpler approach to this problem.

The Bullshit Alarm

My mental model on this is a work in progress. It seems to work, but betting against AI is usually pretty safe when it sounds too good to be true. Recently, Anthropic tried to hype up Mythos, only to be humiliated. Robert Hart, from The Verge, also said:

Anthropic has a habit of using dramatic, alarming-sounding language that can be tough to interrogate cleanly, including flirting with the idea that its Claude model might be conscious.

If a headline is hyperbolic in its implied drama, then my Bullshit Alarm goes off. I’m sure this is the same for anybody taking the time to read this post, or anyone who is terminally online. Anthropic isn’t the only company guilty of alarm-sounding. And we can also blame the journalists and their parent sites for some of this drama. They are just doing their jobs, but seemingly very often taking a quote and spinning it into something more. You may have noticed five of the linked stories are from The Verge, which is owned by Vox Media, one of the leading publishing companies¹ vying for your clicks and attention today. On a slow news day, it apparently pays to post about AI because certain posts read like sponsored content:

• Claude is connecting directly to your personal apps like Spotify, Uber Eats, and TurboTax (The Verge)

Bias!

You don’t like AI so of course you read every headline negatively!

Yes. That statement is true. As I had to explain to a colleague recently, I do not trust the companies that make these LLMs. Because of this, I am unable to speak in a neutral or objective manner on the topic. We are a few years into this nonsense, and this so-called tool has yet to live up to the hype.

We’ve seen layoffs:

• Meta is downsizing by about 10 percent (Endgadget)
• Layoffs.fyi

These should save companies money, yet the savings are not passed on to paid users or consumers. There are expensive data centre projects which create temporary employment during the construction phase. And, did anyone ever become a “prompt engineer” earning six figures?

It’s too soon to tell how big prompt engineering will become, but a range of companies and industries are beginning to recruit for these positions. Anthropic, a Google-backed AI startup, is advertising salaries up to $335,000 for a “Prompt Engineer and Librarian” in San Francisco. Applicants must “have a creative hacker spirit and love solving puzzles,” the listing states. Automated document reviewer Klarity is offering as much as $230,000 for a machine learning engineer who can “prompt and understand how to produce the best output” from AI tools.

The AI Job That Pays Up to $335K — and You Don’t Need a Computer Engineering Background (Time)

Fast-forward 2 years…

“Two years ago, everybody said, ‘Oh, I think prompt engineer is going to be the hot job.’” said Jared Spataro, chief marketing officer of AI at Work at Microsoft. “It’s not turning out to be true at all.”

The Hottest AI Job of 2023 Is Already Obsolete (The Wall Street Journal)

I am biased, and I do live in an echo-chamber. I believe that companies that claim such greatness should stand up to the scrutiny of a middle-aged slacker.

Humbugs

humbug (n.)

Something intended to deceive; a hoax or fraud.

A person who claims to be other than what he or she is; an impostor.

Nonsense; rubbish.

“I never grant favors without some return,” said Oz; “but this much I will promise. If you will kill for me the Wicked Witch of the West I will bestow upon you a great many brains, and such good brains that you will be the wisest man in all the Land of Oz.”

L. Frank Baum, The Wonderful Wizard of Oz

An important aspect about modern technology, especially any technology that claims to be worth billions, is that we cannot see behind the screen. We aren’t even aware there is a screen. Demonstrations are reserved for VIPs in controlled environments. OpenAI, Anthropic, Mistral, they ask us the favour of our patience and promise to bestow greatness upon us. We should be wary.

Toto jumped away from him in alarm and tipped over the screen that stood in a corner. As it fell with a crash they looked that way, and the next moment all of them were filled with wonder. For they saw, standing in just the spot the screen had hidden, a little, old man, with a bald head and a wrinkled face, who seemed to be as much surprised as they were.

Id.

When trying to wrap our heads around the deluge of AI news, it is reasonable to be doubtful. Not about yourself, though. But rather of the company and their promises. Being on the cover of a magazine (like Elizabeth Holmes below) or making Time magazine’s 100 most Influential Companies list doesn’t mean a thing.

“Making believe!” cried Dorothy. “Are you not a great Wizard?”

“Hush, my dear,” he said; “don’t speak so loud, or you will be overheard — and I should be ruined. I’m supposed to be a Great Wizard.”

“And aren’t you?” she asked.

“Not a bit of it, my dear; I’m just a common man.”

“You’re more than that,” said the Scarecrow, in a grieved tone; “you’re a humbug.”

“Exactly so!” declared the little man, rubbing his hands together as if it pleased him; “I am a humbug.”

Id.

Closing

Returning to the initial question, how can we navigate the sea of AI headlines? You can spend your time dissecting, or you could choose to believe it when you see it. Scepticism and doubt are logical reactions. If you think your job could be affected, then you need to ask a different question about standing up to HIPPOs (see What Happens When a ‘HiPPO’ Runs Your Company?) instead of letting the fear get to you.

Remind yourself that newspapers need your clicks to survive. If you want to see fewer stories about AI, don’t click on the stories about AI. It isn’t much, but it is something. Don’t trust the CEOs. Zuckerberg promised the metaverse so hard that they rebranded the company to Meta, and now they are shutting it down. AI companies are playing Oz asking us to wait and investors to invest, then kicking the can down the road. You can make whatever you want of that fact.

Finally, if you are absolutely fed up with reading about AI or hearing about it, just ignore it. Close the webpage, change the channel. If something truly important and groundbreaking happens, it will be talked about enough that you’ll hear about it.