In engineering you learn from these mistakes and try never making them again. Do we want to go through this evolution every time we solve a software issue? Just because we can with an unlimited number of cheap tokens? I think not, I’d rather use the knowledge build up that also knows about the edge cases forgotten to test. Or better, use multiple models that evaluate each other, as Entropic describes it in their recent report https://news.ycombinator.com/item?id=49316271

> We expect that agents coordinating in the wild will act in higher variance ways than we see here, because they’ll have different backgrounds and therefore different contexts. They also, presumably, won’t all be Claudes.

                    Intel    Cost    
                     lig      per
                     ence    Task   Speed

  Gemini 3.7 flash    56     0.40    338
  GLM 5.3 flash       57     0.09     49
  Factor               1      4.4    6.9

> We expect that agents coordinating in the wild will act in higher variance ways than we see here, because they’ll have different backgrounds and therefore different contexts. They also, presumably, won’t all be Claudes.

3 comments https://news.ycombinator.com/item?id=48734656

  ELI5 = Explain Like I'm Five

It seems the 25kV https://en.wikipedia.org/wiki/Betuweroute is missing. Although mostly used for freight and therefor no or other type of schedule, it would be nice if it was projected at least.

See https://news.ycombinator.com/item?id=48883538 25 days ago

> The Sleev (the project has been renamed to make a startup) creator was shilling their project in the OpenCode Discord. That person is very convinced they have something that no one has ever built before. They focused on token reduction without any real evals for capability impacts.

I'm generally against this context pruning without prompting or details. Sleev is very opaque about how it works and definitely will bust your cache.

Dutch https://archive.is/fyCHw

Discussed at The software group Palantir paid just £2M in corporation tax in the UK in 2024 https://news.ycombinator.com/item?id=49180535

OpenRouter doesn’t list DeepSeek as ZDR combined with Caching. Thanks for the tip of Fireworks as Fireworks has Zero Data Retention by default. [1]

[0] https://openrouter.ai/docs/guides/features/zdr

[1] https://docs.fireworks.ai/guides/security_compliance/data_ha...

[0] https://primeradiant.com/blog/2026/superpowers-6.html

Source?

Not withstanding that people are not willing to make sacrifices, my impression is that the last decades young women gained the freedom of enjoying sex as much as they like not restricted by religious or other norms. Sleeping with several different partners a week is not frowned upon but encouraged by friends & family ’Good for you, enjoy!’.

And related to the discussion: safe sex with contraception imho is among the most environmental friendly activities as long as one travels by foot or bike.

You mean no portrait?

  DeepSeek V4 Pro    0.83%
  DeepSeek V4 Flash  2%

[0] https://news.ycombinator.com/item?id=48317294#48317823 [1] https://openrouter.ai/docs/guides/features/response-caching#...

> Apple would have to give any virtual assistant direct access to users’ private data — and the ability to directly control other installed applications — as soon as Siri AI is made available in the EU

At the moment of writing https://news.ycombinator.com/item?id=48343690 MiMo V2.5 Pro had a lower cache hit ratio. From the article:

OSS models, depending on who you use them from, make a huge difference, mostly due to cache-hit rates.

  Model                   Cheapest effectiveInputPrice (Provider)  
  MiMo-V2.5-Pro           0.3720 (Xiaomi) 
  DeepSeek V4 Pro (Max)   0.0560 (DeepSeek)