Hacker News: lordgrenville

Deciphering basmala

lordgrenville — Wed, 24 Jun 2026 09:05:46 +0000

Article URL: https://blog.plover.com/lang/bismillah.html

Comments URL: https://news.ycombinator.com/item?id=48657131

Points: 85

# Comments: 33

New comment by lordgrenville in "Elevated error rate across multiple models"

lordgrenville — Tue, 23 Jun 2026 18:29:10 +0000

There was a post on here the other day explaining this exact phenomenon: https://brooker.co.za/blog/2026/06/19/waiting.html

New comment by lordgrenville in "SpaceX to buy Cursor for $60B"

lordgrenville — Wed, 17 Jun 2026 12:41:28 +0000

> Many made the same prediction about Twitter, and it seems to be more or less the same or higher activity than before, and Bluesky is continuing to rapidly decline

Not at all correct, they are plummeting.

"In June 2025...X year-over-year growth declined by 15.2%." https://techcrunch.com/2025/07/07/threads-is-nearing-xs-dail...

New comment by lordgrenville in "Don't trust large context windows"

lordgrenville — Sun, 14 Jun 2026 11:52:53 +0000

Almost every comment here is appealing to personal experience. By contrast, OP refers to two studies that compare performance on some kind of standardised test over a range of models.

Can't speak to how good those tests are, but they can't be worse than anecdotal evidence for something as vague/subjective as LLM performance.

New comment by lordgrenville in "Cleaning up after AI rockstar developers"

lordgrenville — Tue, 09 Jun 2026 18:09:45 +0000

I've never heard it before, but immediately understood it and found it a useful term for something I didn't have a single word for.

New comment by lordgrenville in "The newest Instagram “exploit” is the goofiest I've seen"

lordgrenville — Tue, 02 Jun 2026 12:01:02 +0000

It's not really an attack surface though. Reminds me of https://xkcd.com/932/

New comment by lordgrenville in "Magnifica Humanitas"

lordgrenville — Mon, 25 May 2026 11:55:09 +0000

"But the effect of her being on those around her was incalculably diffusive: for the growing good of the world is partly dependent on unhistoric acts; and that things are not so ill with you and me as they might have been is half owing to the number who lived faithfully a hidden life, and rest in unvisited tombs."

New comment by lordgrenville in "The river otter's remarkable comeback"

lordgrenville — Sun, 10 May 2026 18:28:41 +0000

I recently read Tarka the Otter [1], which is the story of a life of an otter. At the end he is hunted by a farmer and a pack of otter hounds. It is pretty brutal. Reminds you that at the time they were seen as vermin and a nuisance, not cute and furry.

[1] https://www.gutenberg.org/ebooks/75899

Rotten Dot Com

lordgrenville — Sun, 10 May 2026 08:25:55 +0000

Article URL: https://www.theparisreview.org/blog/2026/05/06/rotten-dot-com/

Comments URL: https://news.ycombinator.com/item?id=48082039

Points: 135

# Comments: 126

New comment by lordgrenville in "Agents for financial services and insurance"

lordgrenville — Wed, 06 May 2026 10:26:17 +0000

I think OP meant that Farmville was built by a different party (Zynga). FB was trying to encourage other businesses to build apps on its platform, not build them itself.

New comment by lordgrenville in "He asked AI to count carbs 27000 times. It couldn't give the same answer twice"

lordgrenville — Wed, 29 Apr 2026 13:16:40 +0000

Wow, yeah. "The result is an app that the creators say is 90% accurate".

https://techcrunch.com/2025/03/16/photo-calorie-app-cal-ai-d...

New comment by lordgrenville in "We still don't have a more precise value for "Big G""

lordgrenville — Wed, 29 Apr 2026 06:06:24 +0000

Are you sure these are actually issues, or are you just trusting that the AI criticism is correct?

New comment by lordgrenville in "Show HN: A free ESG stock screener that publishes its losses and methodology"

lordgrenville — Sun, 26 Apr 2026 13:19:53 +0000

As Matt Levine often points out, there are two possible cases for ESG

1/ This will bring worse returns, but I'm willing to accept the loss in order to forward values I support

2/ This will bring better returns, since the market underrates risks from bad ESG companies (e.g. the long-term return on capital for coal companies will be worse than the market expects)

People marketing ESG funds (or anti-ESG, same rule applies) usually emphasise the second.

> Anyone claiming they can consistently beat any large index is just delusional, aren't they?

This is obviously not true. RenTech would like a word.

New comment by lordgrenville in "GitHub's fake star economy"

lordgrenville — Mon, 20 Apr 2026 12:24:40 +0000

More to the point, in the US losing teams get rewarded in the form of draft picks, which sometimes creates perverse incentives. This doesn't exist in European football. (Disclaimer: I know almost nothing about American sports.)

New comment by lordgrenville in "Advancing human gut microbiota research by considering gut transit time (2023)"

lordgrenville — Mon, 20 Apr 2026 07:55:43 +0000

n=1

But interesting nonetheless, thanks for sharing your findings.

New comment by lordgrenville in "How to make buffet breakfasts less wasteful"

lordgrenville — Fri, 17 Apr 2026 09:22:24 +0000

I was surprised that this article is about food wasted by people not finishing their plates. Would have guessed that a lot of the unserved food is discarded (sure, some of it can be served at tomorrow's breakfast, but only within limits), and that this is much more significant.

New comment by lordgrenville in "Google's 200M-parameter time-series foundation model with 16k context"

lordgrenville — Tue, 31 Mar 2026 08:38:31 +0000

That's what traditional time-series modelling does. This is a foundational model, which means it's just a neural network trained on lots of time series. (So maybe OP's question still stands? But it's the same question as "how can LLMs be good at so many different kinds of conversations?")

New comment by lordgrenville in "CodingFont: A game to help you pick a coding font"

lordgrenville — Mon, 30 Mar 2026 18:53:04 +0000

Nice idea, would be good to add a third option for "these look indistinguishable" (and then I guess they could be bundled together in later stages).

New comment by lordgrenville in "Meta acquires Moltbook"

lordgrenville — Wed, 11 Mar 2026 13:25:54 +0000

I expect that OP just meant "native language"

New comment by lordgrenville in "Tell HN: YC companies scrape GitHub activity, send spam emails to users"

lordgrenville — Thu, 26 Feb 2026 15:19:18 +0000

Fair point. Pretty sure there is a way to have a few .gitconfig files, with the active one based on the remote URL domain, but it is more work.