ok, so I asked the AI to have a look at this and asked the question, what if the original email was written by an AI (I hope you don't mind reading the AI output)

You’re describing Google/Gmail adding an AI layer that: reads all your emails, extracts “key topics,” infers obligations (“you should do X”), and presents a distilled action list.

That’s already risky. But the real problem appears when the input itself was AI-generated. The AI AI compression loop (this is the core issue)

Here’s the pipeline that will increasingly exist in workplaces:

Human intent

  LLM writes email (already compressed, polished, “helpful”)

  Recipient’s AI reads that email

  AI summarizes, extracts tasks, infers priorities

  Human acts on the summary, not the original

Each step removes information.

Crucially:

No step knows what was removed by the previous step.

This is not neutral loss. It’s directional distortion.

Why this is uniquely dangerous

1. LLM-written emails already discard uncertainty

When an LLM writes an email, it tends to: smooth ambiguity, resolve hedging, strengthen modality (“should” “need to”), infer intent that the human didn’t explicitly commit to.

So the original email is already an interpretation, not a raw signal.

2. The summarizing AI treats that interpretation as ground truth

The inbox AI then: assumes the email is authoritative, extracts obligations as if they were explicit, converts soft language into tasks, removes qualifiers as “noise.”

Now you have: Inferred intent being treated as fact.

3. Humans stop reading originals

This is the most important part.

In practice: people will skim the AI Inbox, trust the to-do list, and only open emails if something looks wrong.

But if both layers are AI-generated, nothing looks wrong — it looks clean.

That’s how errors propagate silently.

Concrete workplace failure modes

Here’s what will actually happen:

  Obligations get invented

“Let’s discuss rescheduling”
  AI email writes: “We should reschedule”
  Inbox AI says: “Reschedule meeting”

No one ever decided to reschedule.

  Responsibility shifts invisibly

LLM phrases something politely.
Inbox AI converts it into your task.
Now you’re on the hook.

  Legal and compliance risk explodes

In regulated environments: the AI summary becomes the de facto record, not the original wording, even though the summary has no legal standing.
This is catastrophic for: finance, law, immigration,HR, contracts.

As the first commenter wrote "AI is not the first thing to come along that wiped out whole categories of jobs."

What's special? Imagine that in any office that you walk into, a minimum of 20% get made redundant within the next 3 years.

I work in marketing, here now, you don't need agents, you just need on person who is capable - an editor that can read and has taste. Get that person to prompt and edit manually the AI slop. That's 3 or 4 jobs now done by one person. Copywriters and proofreaders are not longer needed. Just keep the editor.

If you think it's hard to find a job now? If you think that employers have too much power of employees now? Wait 'til you see what's coming.

Editor/Copywriter. "Good enough" is perfect when it'll only get a fraction of second of brain time. LLMs are already laying waste to marketing departments around the world. It'll write "all" the product copy, just copypasta the user manual with some additional info, like pricing, sales promotions, etc.

Anyone still saying the gains are negligible, has their head firmly planted in the sand, imo. Jobs far below your pay-grade aren't about to be decimated (1/10), no, more hung and quartered (4/10). Anything involving marketing, sales, writing and storytelling in general, a lot less people will be able to make a living from this.

There's going to be a hell of a lot of people out of work soon. We're heading into a employers market, a long one. For the first time in my life, I'm feeling the "prepper" vibe. "Shit's about to get very real"...but they've been saying that forever, haven't they?

Wrong fucking game you idiot.

You start with this "Though it's been years since I played TF2, I just love that game." and then continue with "and people started building walls everywhere so the maps were useless and unplayable. The teleporting cheats were a massive downer as well."

That's fortnight you bell-end, not TF2.

I've been playing TF2 since 2011. It's never been a better time to play! Great new maps getting dropped once or twice a year recently. Sadly, no hoov update yet.

I own 5 and 6. I still play five.

Six was so boring, it felt like it was a office job.
Another thing, I never really got past the awful graphics, the weird-looking characters with like 1000 options on how to respond.
Civ V had a tech tree. Civ 6 had multiple trees.

Make a game so complicated that it's boring to play.

Civ V, is still so playable. So much so that the last time I looked, it was more expensive on STEAM than VI.

I'm not the only one here thinking this.
FYI, I stared with Civilization, (1) the first.

With that in mind, I'll wait 5 or 6 years until all of the DLC for 7 are completed and all of the reviews are in, before I sink 100 hours into a boring game waiting for it to be fun. (I've been playing V, on and off for more than a decade.)

But I guess 10,000 hours of TF2 says differently.

"There is no chance they were chosen on quality and every chance they were picked to serve some personal agendas."

The Hugos became irrelevant when the 5th Season won over Seveneves.

MS, while we're here fixing Edge, could you please have a look at how your browser cooperates with ChatGPT.

Basically, it won't work at all. Chrome is no better. But, oh boy, ChatGPT just loves Firefox.

Funny that. I use chrome for work, exclusively. Edge for nothing, so I wanted to sandbox the AI here. Firefox has been my main personal browser (think porn, drugs, shit posting) since reading a story on the browser on /. back in the early 2000s.\

Am i missing something about the data ChatGPT is scraping?

"China explicitly promises not to nuke anyone who doesn't nuke them first"

The Chinese government isn't trustworthy in any sense. Absolutely nothing said can be believed. In fact, to strike first then claim that it was in self defense would be very "China."

But with Russia, we had people to talk to. When we called, someone picked up. In the new China (and Russia) power is so concentrated at the top that (my guess) generals and politicians are afraid to pick up the phone unless the leader has already formulated a plan.

Avoiding a war only works if there's someone to talk to. Hell, the new China head of the army is under US sanctions. That tells you how much the Chinese want to talk to us.

Holy shit...between star wars and batman...I just don't give a shit anymore. Another cgi crap film about flying people with magical powers. I did watch patterson's 'the batman'. 3 hours long, this was a movie that just didn't need to be made. No purpose. A story that's been flogged to death...it's now...'rebooted.'

Congratulations! And...wow, 27417, you must be old.

Pac-man? I can honestly say I've never played it. Every arcade box I can remember was always Ms. Pac-Man.

no, no, no, it was actually better than the original.

Then maybe it's time to retire some of the old sci-fi. I would much prefer to see Seven Eves made into a movie.

And in book two, Paul and his revolution is responsible for 60 billion deaths. (But actually in Frank's story, without Paul - or actually, his son, 'cos Paul chickens out, we go extinct.)

What kills me is that Herbert's notes for book seven were never released. Instead, we got his son's absolute crap writing. And none of the final questions properly answered (such as where did the spice come from, Arrakis was once a water planet before the worms). IMHO, Frank's universe building trumps Tolkien's black and white (good/evil) story about faries and dwarves.