Hacker News Reader

stefap2on Artificial intelligence now beats some of the best human forecasters

Are the models going to skew their analysis to preserve AI companies as a form of self-preservation?

stefap2on More than 100k people in Japan are now aged 100 or older

One would say a leisurely life in France, but on the other hand, the Japanese are not known to take a leisurely approach to life.

stefap2on More than 100k people in Japan are now aged 100 or older

88% are women.

stefap2on Ask HN: What are you working on? (September 2026)

  year  projects  avg/month  HN posts/day  projects per 100k HN posts
  ----  --------  ---------  ------------  --------------------------
  2008       123       10.2           880                        38.2
  2009       104        8.7         1,669                        17.1
  2010       666       55.5         2,826                        64.6
  2011        86        7.2         3,713                         6.3
  2012        46        3.8         4,315                         2.9
  2013       389       32.4         5,485                        19.4
  2014       146       12.2         5,009                         8.0
  2015       104        8.7         5,476                         5.2
  2016       471       39.2         6,758                        19.0
  2017       456       38.0         7,534                        16.6
  2018       242       20.2         7,546                         8.8
  2019       246       20.5         8,569                         7.9
  2020       882       73.5        10,036                        24.0
  2021       837       69.8        11,384                        20.1
  2022       363       30.2        12,187                         8.2
  2023       425       35.4        12,675                         9.2
  2024      2122      176.8        10,203                        56.8
  2025      6220      518.3        10,649                       160.0
  2026      6093      677.0        12,641                       187.5  (Jan-Sep)

stefap2on Ask HN: What are you working on? (September 2026)

projects per year in "Ask HN: What are you working on?"

  year  projects  avg/month
  ----  --------  ---------
  2008       123       10.2
  2009       104        8.7
  2010       666       55.5
  2011        86        7.2
  2012        46        3.8
  2013       389       32.4
  2014       146       12.2
  2015       104        8.7
  2016       471       39.2
  2017       456       38.0
  2018       242       20.2
  2019       246       20.5
  2020       882       73.5
  2021       837       69.8
  2022       363       30.2
  2023       425       35.4
  2024      2122      176.8
  2025      6220      518.3
  2026      6093      677.0  (Jan-Sep)

Top five topics per year (% of that year's projects, keyword-classified):

  year  #1             #2             #3             #4             #5
  ----  -------------  -------------  -------------  -------------  -------------
  2008  Dev tools  9%  Games      9%  Video/gfx  7%  Writing    7%  Business   6%
  2009  Dev tools 16%  Science   11%  Writing    8%  Education  7%  Business   7%
  2010  Dev tools  8%  Games      6%  Video/gfx  6%  Business   4%  Writing    4%
  2011  Dev tools  8%  Games      8%  Social     7%  Business   7%  Science    7%
  2012  Business  13%  Dev tools  7%  Hardware   4%  Video/gfx  4%  Social     4%
  2013  Dev tools 16%  Video/gfx 10%  Games      7%  Writing    6%  Education  6%
  2014  Dev tools 12%  Video/gfx 10%  Writing    7%  Hardware   6%  Games      6%
  2015  Games     11%  Dev tools 10%  Business  10%  Hardware   8%  Video/gfx  8%
  2016  Dev tools 17%  Video/gfx 11%  Hardware   8%  Business   6%  Music      5%
  2017  Dev tools 18%  Video/gfx 10%  Hardware   7%  Games      6%  Data       5%
  2018  Dev tools 14%  Video/gfx  8%  Writing    6%  Games      5%  AI/LLM     4%
  2019  Dev tools 17%  Video/gfx 10%  Games      9%  Hardware   6%  Business   5%
  2020  Dev tools 14%  Video/gfx  9%  Games      8%  Hardware   5%  Writing    5%
  2021  Dev tools 18%  Video/gfx 10%  Games      9%  Hardware   7%  Writing    6%
  2022  Dev tools 21%  Video/gfx 13%  Music      9%  Science    8%  Games      7%
  2023  Dev tools 14%  Video/gfx 10%  Games      9%  Writing    8%  AI/LLM     7%
  2024  Dev tools 20%  AI/LLM    17%  Video/gfx 11%  Games      9%  Business   6%
  2025  AI/LLM    27%  Dev tools 21%  Video/gfx 10%  Games      9%  Data       5%
  2026  AI/LLM    33%  Dev tools 22%  Video/gfx 10%  Games     10%  Hardware   6%

stefap2on On the Navier–Stokes Millennium Prize Problem

Maybe I should have been clearer. My point is that solving something like Navier–Stokes just pushes knowledge work further ahead, onto a new set of bigger and more complex problems. Navier–Stokes is a Millennium problem today, but once problems like that become solvable, they can open the door to entirely new classes of problems we haven’t even thought of yet.

stefap2on On the Navier–Stokes Millennium Prize Problem

Wow, this sentence is doing a lot of work in that tweet: "While unlikely, we cannot rule out that de-identified data derived from their usage of our products helped improve our models."

stefap2on On the Navier–Stokes Millennium Prize Problem

This just pushes knowledge work further up the ladder, toward larger and more complex problems. If there are no knowledge workers, who is going to interpret these results, validate them, decide what matters, and put them into practical use? Rather than eliminating knowledge work, advances like this could create entirely new layers of problems to solve and opportunities to pursue, which will create even more jobs and opportunities. This is my optimistic take.

stefap2on We Tested Nonstick Cookware: Coatings Don't Need to Look Worn to Shed Particles

Where does the hexclad fall in this regard?

stefap2on Willingness to look stupid

I found it gets easier as you get older. Somehow I care much less what others think

stefap2on Don't post generated/AI-edited comments. HN is for conversation between humans

If it takes longer to read, it's not an AI problem, but the author failing to catch that the comment is too drawn out. I don't see how it is a problem to have AI write a comment if you agree with the content. If it is bad content, it will eventually reflect badly on the author anyway.

stefap2on Amazon closing its Fresh and Go stores

And Amazon also discontinuing Amazon One palm authentication services in whole foods. I wonder if these are related events. https://news.ycombinator.com/item?id=46790734

stefap2on Show HN: I quit coding years ago. AI brought me back

The site is down.

stefap2on Influencers and OnlyFans models are dominating U.S. O-1 visa requests

Actually, it falls under OF-1 visa category.

stefap2on Everything's Fake Now – Liam Fallen

Grammar seems to have improved on Instagram posts :)

stefap2on Ask HN: AI has changed my job for the worse

I read what you say, but I can hardly imagine how people can be twice as productive. I use Claude daily for writing various scripts. It produces a lot of code, so yes, if I measured it by generated code, then it's productive. However, it works best for short, clearly defined tasks/scripts. If you want anything longer, you have to watch it like a hawk so it doesn't go off track. And I'm not sure if I would give this code to someone else to use without a thorough review. I would say it may be even less productive (I think there was a paper on that).

I also use LLMs for writing; it's good, but again, you need to carefully read everything and rewrite passages that are completely made up. So I'm not really sure how this can replace people, to the point that Amazon is firing 30,000 people. I have a hard time squaring it that it's because of AI.

stefap2on AI assistants make widespread errors about the news, new research shows

Some 72% of responses by Gemini, Google's AI assistant, had significant sourcing issues, compared to below 25% for all other assistants, it said.

Issues of accuracy were found in 20% of responses from all AI assistants studied

stefap2on AI is making us work more

When developers are allowed to use AI tools, they take 19% longer to complete issues—a significant slowdown that goes against developer beliefs and expert forecasts. This gap between perception and reality is striking: developers expected AI to speed them up by 24%, and even after experiencing the slowdown, they still believed AI had sped them up by 20%.

https://news.ycombinator.com/item?id=45661775

stefap2on Measuring the Impact of Early-2025 AI on Experienced Developer Productivity

stefap2on Largest Mass Resignation in US History as 100k Federal Workers Quit

What i have seen is that those who resigned had 20+ years experience and younger folks stayed. So it was a bit of a brain drain, at least from my limited interaction with the three letter agencies.

stefap2on How AI slop is clogging your brain

"And there's something really freeing about taking back some of your attention and time and just walking around unmonetized."

stefap2on Amazon to end commingling after years of complaints from brands and sellers

Can you give some examples? I order massive amounts from Amazon and I don't think I have received any counterfeit items. Most of it is made in China.

stefap2on Amazon to end commingling after years of complaints from brands and sellers

I’m more irritated by the comingling of small, fragile products with large, heavy ones in the same box, like a box of crackers shipped together with detergent. Because of this, I had to stop my Subscribe & Save orders. Every month they would do this nonsense.

stefap2on Scammed out of $130K via fake Google call, spoofed Google email and auth sync

I have a cell phone with an area code where I no longer have any connections or ties. Almost all the spam calls I receive come from that area code. By simply ignoring or blocking calls from that area code, I can avoid nearly all of the spam.

stefap2on Donald Trump's fantasy of home-grown chipmaking

Did they blow it? From a shareholder’s perspective, if you don't have ideas on how to increase shareholder profits through innovation, then buying back shares is the right thing to do.

stefap2on Cognitive decline can be slowed down with lifestyle changes

Structured lifestyle intervention. Participants attended 38 facilitated peer team meetings over two years and were provided with a prescribed activity program with measurable goals for:

- aerobic, resistance and stretching exercise

- adherence to the MIND diet, a combination of the Mediterranean diet with the salt restrictions of the DASH diet

- cognitive challenge through BrainHQ training and other intellectual and social activities

- regular review of health metrics and goal-setting with a study clinician

stefap2on Telo MT1

I thought electric cars are simpler and should be cheaper. You can tariff out competition only for so long...

stefap2on Trump Is Losing Patience with Musk's Outbursts over Megabill

Musk will alienate both dems and reps, who is going buy teslas:)

stefap2on Getting AI to write good SQL

I have done this using the OpenAI 4o model. I had to pass in a prompt with business-specific instructions, industry jargon, and descriptions of tables, including foreign keys. Then it would generate even complex join queries and return data. In my case, I was more interested in providing results to users not knowledgeable about SQL, but the SQL was displayed for information.

stefap2on Perverse incentives of vibe coding

I found useful asking it to write with minimal lines. Let it write a functional code and ask to remove all unnecessary lines, minimalistic, efficient code.