Matthew Hughes
@matthewhughes.bsky.social
1.7K followers 1.6K following 1K posts
Journalist (ex The Register, The Next Web, HowToGeek). Writer. Software developer. Dog owner (x3). Scouser. I have a newsletter about how tech companies are ruining our lives. https://whatwelost.substack.com/
Posts Media Videos Starter Packs
Pinned
matthewhughes.bsky.social
My latest newsletter is about how the generative AI industry made us fear the wrong things.

AGI won't kill us all, and AI won't take our jobs. But it'll still harm us.

Generative AI a paper tiger, with a real tiger behind it.

whatwelost.substack.com/p/generative...
Generative AI is a Paper Tiger with a Real Tiger Behind It
We Were Afraid Of The Wrong Stuff
whatwelost.substack.com
matthewhughes.bsky.social
It’s genuinely funny getting research/analyst quotes for @edzitron.com and seeing how people react when I say my questions are for Ed.
matthewhughes.bsky.social
That’s such a good headline though.
Reposted by Matthew Hughes
shane.co.uk
The first "UK Porn Block" / "Child Safety" data breach is out.

https://www.theguardian.com/games/2025/oct/07/discord-data-breach-proof-of-age-id-leake
d

I knew it would be quick, but not that quick.
matthewhughes.bsky.social
One of my potentially unpopular opinions is that getting a first aid certificate should be a prerequisite to hold a driver’s license, as it is in Germany.

Also, well done!
matthewhughes.bsky.social
It’s just spectacular to see how committed Microsoft is to killing its game business.
matthewhughes.bsky.social
Also that (and I say as much in the piece.
matthewhughes.bsky.social
Oh, it’s more along the lines of “I want to support these people.”

I have genuinely paid extra to buy genuine Arduino in the past, even though it costs 10x as much (more if bought from Maplin, RIP), because I wanted to support them.

Even though legally there’s nothing wrong with the knock-offs.
Reposted by Matthew Hughes
edzitron.com
I'll be joining @benjedwards.com at 3:30 ET to talk about the AI bubble, join us!
benjedwards.com
Heads-up that I’ll be talking to @edzitron.com about the AI hype machine and the “AI bubble” this afternoon live on YouTube: 3:30 P Eastern, 12:30 P Pacific

arstechnica.com/ai/2025/10/a...
Ars Live: Is the AI bubble about to pop? A live chat with Ed Zitron.
Join a live discussion on October 7 about the AI gold rush.
arstechnica.com
Reposted by Matthew Hughes
jonchristian.net
folks, what are we even doing here
In one video, a forklift delivers Hawking’s wheelchair into a WWE-style ring, where he’s immediately knocked to the ground by burly wrestlers.

“This shouldn’t even be legal!” exclaims an announcer in the AI-generated clip.

In another video, Hawking takes blow after blow to the face from a UFC fighter. “Hawking’s in trouble,” the announcer yells, as the physicist topples out of his wheelchair.

In another, Hawking is trampled by a raging bull. In another, he’s attacked by a crocodile that drags him into a muddy river. In yet another, he drops into a skateboarding halfpipe and wipes out badly, smoke pouring from his wheelchair.
matthewhughes.bsky.social
Those risks are the real tigers.

These are the things we need to be wary of. The things we need to fight back against. And the things that we, like Paul Revere, need to warn the world about.

whatwelost.substack.com/p/generative...
Generative AI is a Paper Tiger with a Real Tiger Behind It
We Were Afraid Of The Wrong Stuff
whatwelost.substack.com
matthewhughes.bsky.social
The biggest lie the generative AI industry ever told is that it was a threat to our livelihood and our lives.

This disguised other, and arguably more sinister threats, and as a result, we're oblivious to the risk that this technology actually poses.

whatwelost.substack.com/p/generative...
matthewhughes.bsky.social
Instead, we're seeing short-form videos generated with Veo3 and Sora that don't pretend to be anything other than AI-generated, and that try to reinforce an idea through tropes and low-brow jokes.

A lie, repeated often enough, becomes truth.

whatwelost.substack.com/p/generative...
matthewhughes.bsky.social
Even when it comes to disinformation, generative AI's threat isn't what we were first warned about.

We expected sophisticated deepfakes, or fake news publications churning out copy at a scale faster than the New York Times.

whatwelost.substack.com/p/generative...
matthewhughes.bsky.social
This is the real tiger hiding behind generative AI.

ChatGPT was never the problem. It was the Jack Welch clones that helm every FTSE 100 and Fortune 500 company.

whatwelost.substack.com/p/generative...
matthewhughes.bsky.social
If it was anything else, why would Meta bring back stack ranking, and start purging "low performers"?

Why are some of the WITCH "big five" outsourcing companies seeing faster revenue growth than Apple, and all but one having positive growth?

whatwelost.substack.com/p/generative...
matthewhughes.bsky.social
Three years into this farce, there is no evidence that AI is taking people's jobs, as a recent Yale study shows.

The tech job market sucks, but that's because companies are taking advantage of the moment to outsource and cutback on pandemic-era headcount.

whatwelost.substack.com/p/generative...
matthewhughes.bsky.social
AI, they said, would kill the market for software engineers.

Meanwhile, some software engineers are making bank fixing vibe-coded apps, in part because coding != software engineering.

whatwelost.substack.com/p/generative...
matthewhughes.bsky.social
They said agents -- not people -- were the future.

The same agents that don't work, and that the people who are forced to use them absolutely despise.

My favorite review of Salesforce's Agentforce: "Using Agentforce is like vomiting on a piece of shit."

whatwelost.substack.com/p/generative...
matthewhughes.bsky.social
It's important to understand what generative AI is actually doing, and compare it to what we were warned against.

We were told that AI would destroy entry-level jobs within a few years. We were told that genAI models wouldn't hallucinate by 2025.

whatwelost.substack.com/p/generative...
matthewhughes.bsky.social
With generative AI -- and AI more broadly -- we were warned that it might destroy the job market, and perhaps also humanity.

The only other technology that I could think of that parallels this is...

Wait for it.

The nuclear bomb.

whatwelost.substack.com/p/generative...
matthewhughes.bsky.social
As a society, we're really bad about predicting the negative impacts of new technological breakthroughs. And so, when something new bursts on the scene, it's all positive. The bad shit comes later.

Generative AI was different, however.

whatwelost.substack.com/p/generative...