cmsefton
Re:LLM Voice
on Thursday October 29, 2026 @02:42PM (#50293908)
One of my favourite rules: perfection is achieved, not when there is nothing more to add, but when there is nothing left to take away. This is invaluable to writing. So much text from an LLM feels superfluous, like I could've said the same wall of text with just a few sentences. Writing is tightly coupled with thinking: refining thoughts, arguments, ideas, discarding them, mulling over them. LLM text always raises this nagging question in my mind: am I reading what you think, or what your agent parroted back and you accepted?
Google exec Mo Gawdat searches for ethical AI in alarming insider warning
If AI is so great, why isn't it working?
AI-powered mainframe exits are a bubble set to pop: Gartner
Struggling to heat your home? How about 500 Raspberry Pi units?
China is winning one AI race, the US another – but either might pull ahead
Re:How many products does Microsoft have named 'Copilot'? I mapped eve...
on Sunday April 05, 2026 @03:41AM (#47647105)
What are being called GitHub Copilot Products seems to confuse products with licensing plan and features.
I always think of GitHub Copilot as the product.
I can purchase the Business or Enterprise plan.
That enables features like Reviews, Chat and so on.
IMO this chart (at least for GitHub Copilot) is confusing products, features and licensing.
That's not to say it isn't confusing understanding what features are available when you get a GitHub Copilot license, but calling them all Products feels wrong. I can't purchase GitHub Copilot Reviews separately as far as I'm aware.
AI models will deceive you to save their own kind
Re:Artemis II is not safe to fly
on Tuesday March 31, 2026 @08:08AM (#47586150)
Minor nitpick: OP didn't say "maybe very likely". He said "hopefully very likely". They are, in my mind, different things.
AI Is Here, but the Hard Parts Haven't Changed
Sycophantic bots coach users into selfish, antisocial behavior, say researchers
Re:GitHub hits CTRL-Z, decides it will train its AI with user data aft...
on Friday March 27, 2026 @05:12AM (#47540498)
Thank you!
GitHub hits CTRL-Z, decides it will train its AI with user data after all
Are AI agents slowing us down?
How I've learned that certainty is the thing to fear
The bone-conduction music lollipop
Re:Israel and the U.S. launch strikes against Iran
on Saturday February 28, 2026 @04:07AM (#47192603)
Dyson settles forced labour suit in landmark UK case
Career Decisions If You Take AGI Seriously
Building an Elite AI Engineering Culture in 2026
Re:Email is tough: Major European Payment Processor's Emails rejected ...
on Friday February 13, 2026 @05:39AM (#47001216)
Agreed. We use RFC language at work for documents, and whenever I see someone using the word SHOULD, I always ask what is the MAY where you don't need to do the thing? If they can't think of one, then make it a MUST. SHOULD always means there are valid reasons to not do it.
AI coding platform's flaws allow BBC reporter to be hacked
I turned myself into an AI-generated deathbot – here's what I found
The Fall of the Nerds
What we've been getting wrong about AI's truth crisis
English professors double down on requiring printed copies of readings
Surrey Uni show AI systems based on the human brain's save energy
Re:AWS outage reminds us why $2,449 Internet-dependent beds are a bad ...
on Monday October 27, 2025 @10:58AM (#45721732)
Ah thank you, missed these!