Kye

AI writing you can recognize as AI writing is obvious. Newer models are better about this and the line will only get more blurry. Here's a benchmark where good writers make the assessment rather than different LLMs ranking each other: https://surgehq.ai/leaderboards/hemingway-bench

The top models are also the latest:

Gemini 3.1 Pro: still a bit of a gremlin, but will probably stay on top until the other model makers go xkcd 810 and target this benchmark

Gemini 3 Flash: current favorite of writers using it as a helper for its speed and decent prompt following

The rise in prevalence is recent, but older than transformer models by a comfortable margin.

https://books.google.com/ngrams/graph?content=delve&year_sta...

I value people who put in the work. I also value being able to make a little one-off single use gadget without having to spend a week doing remedial python every few months. I can understand it once it's written but writing it is a separate skill.

Of course, having learned a few languages, understanding data types, knowing to prompt it for idiomatic code and check against best practices, etc is vital to being able to do that. The basic skills need to be developed even if not everyone gets the same value out of being able to write code.

While I categorically reject this argument, it would be stronger if you at least attempted to address one or two of the obvious counters like collage.

Linus Tech Tips just did a video on the impact of RAM speed: https://www.youtube.com/watch?v=05HRLuU--jI

This might be one of the applications where it makes a difference.

Who is "you guys"

You're reading your own hangups and bugaboos into my comment. The problem begins and ends with the lack of consent. There are plenty of people who would be into it if he asked first.

You also get 30 years of efficiency improvements and 20 years of capacity and reliability improvements when you replace them.

In practice: https://www.rte.ie/news/regional/2026/0116/1553440-mayo-wind...

>> "Each one of the new wind turbines will be capable of supplying more power to the national electricity grid than was generated by the entire Bellacorick wind farm."

He has a whole video[0] on the difference between energy and electricity, so he understands it. Maybe there's some disconnect between the video and your interpretation.

[0] https://www.youtube.com/watch?v=OOK5xkFijPc

Federation: focuses on diffusing responsibility

P2P: everyone is responsible for their part of it

Atproto: you should be able to choose how much responsibility you actually need or want while mitigating risk through competing stacks and credible exit

Bubble doomerism is nothing novel. As is always the case, he's right vertically and wrong horizontally. Serious people in serious publications still speculated that the internet was a fad and would be over soon as late as 2008.

OpenAI will collapse, almost certainly. Anthropic might get by if they can make it to IPO before it all comes tumbling down. Google will buy up all the datacenters in a fire sale like they did with dark fiber after the .com bubble popped and continue building out stuff like NotebookLM.

Amazon and Microsoft will still be there selling server time to model providers and doing custom enterprise solutions like always. They already host the major proprietary models and sell API access.[0]

The top open models are already good enough. At this point prompting and coordination are the big bottlenecks. It would be nice if the bubble lasts long enough for open models to match at least the latest Opus.

His problem is the focus on the bubble and not on what usually happens after. People will bandy his pieces about insisting it's all short lived and they can just wait it out. Kimi K2.5, GLM 5, and MiniMax 2.5 aren't going away.

[0] For example: https://azure.microsoft.com/en-us/blog/claude-opus-4-6-anthr...

https://aws.amazon.com/about-aws/whats-new/2026/2/claude-opu...

He hedges so much that it's probably impossible to catch him in a contradiction or missed prediction. It must be all that practice running a PR firm for AI companies.

>> "The dataset for this insight combines data on large language model (LLM) API prices and benchmark scores from Artificial Analysis and Epoch AI."

I don't know about Epoch AI, but Artificial Analysis shares its methodology: https://artificialanalysis.ai/methodology

Their chart of inference prices split by benchmark intelligence: https://artificialanalysis.ai/trends#efficiency

He has this thing where he tries to involve people in his ABDL roleplay without consent.

Nobody goes there, it's too crowded.

You could arrive at the essence of this by just having read and internalized Carl Sagan's The Demon-Haunted World. Especially the Baloney Detection Kit.

In my experience good prompting is mostly just good thinking.

Someone with legal responsibility for the data those plugins touch.

See also: ask vs guess culture https://ask.metafilter.com/55153/Whats-the-middle-ground-bet...

An AI product manager agent trained on all the experience of product managers setting budgets for features and holding teams to it. Am I joking? I do not know.

A post I saw the other day from someone in a similar situation who did share what changes were made: https://bsky.app/profile/abumirchi.com/post/3meoqzl5iec2o

I've only submitted a few things, but never seemed to face a selection process. I think the low paywall is the only barrier.

>> "I also don't see why there can't be an open directory of websites where the community makes decisions about what to add instead of leaving it to a single individual."

Because no one who wants one has made it. Why not be the change if it's something you want?

If your response is anything other than enthusiasm to get started, you understand why it hasn't happened.

There's always some friction between implicit assumptions of reader and writer. I assumed they were hand curated. I've never seen algorithmic selection produce the kind of variety I see on there.

>> Another problem is that I like to add a variety of sites so that people following what’s recently added don’t get swamped by loads of blogs on one topic. And last time the site got on HN the suggestions (not “submissions”) were swamped with mostly men with rarely-updated blogs about computers.

Essentially my trouble with every "share your blog" type thing that appears on HN. Some of the blogs do show some interest outside computers, but those posts are quickly swamped by more computer touching.

I appreciate the curation in favor of diversity of interest here.

edit: You can see it in a lot of the suggested alternatives elsewhere. I think it's hard for someone to really get it if computer touching is their life. Curation like this is vital to avoid regression to the mean.

Rolling Stone, for example: https://en.wikipedia.org/wiki/A_Rape_on_Campus#Columbia_Univ...

Which specific use cases does this provide an alternative for? Chat is a tiny part of what people do with Discord and there are plenty of options already.

Try this in ChatGPT: "So ChatGPT is getting ads. The Google guys wrote _the_ paper explaining why ads in search are a bad idea, and Google set about demonstrating it. How can ChatGPT avoid the same fate with all the same incentives?"

"Correct" is doing a lot of work there. Dialects are a thing. I have never heard anyone pronounce it like whee-ah. They would get a lot of chuckles here where it's pronounced the same as where.

Whee-ah is a little emphasis away from sounding like a donkey.

This is probably a good representative take for the view of people who assigned value to the coding part of building software. It's worth reading and saving, if only as an artifact to study.

Sparkle otter

A pelican GIF in a Pelican(TM) MP4 container.