Hacker News Digest

Sunday, April 26, 2026

In This Issue

  • Hacker News
  • 1-Bit Hokusai's "The Great Wave" (2023)
  • The West Forgot How to Make Things. Now It's Forgetting How to Code
  • Trump fires all 24 members of the U.S. National Science Foundation
  • Amateur armed with ChatGPT solves an Erdős problem
  • USB Cheat Sheet
  • The Free Universal Construction Kit
  • The AI Industry Is Discovering That the Public Hates It
  • GPT 5.5 biosafety bounty
  • It's OK to Use Agentic to Revive the Projects You Never Were Going to Finish

Zipper Data Brief

April 26, 2026
Your daily digest of the best from Hacker News

Top 6 Trending

#1
582 points · stephen-hill · comments
Discussion Summary
Commenters celebrate a 1-bit digital recreation of Hokusai's "The Great Wave" created on vintage Mac hardware, praising both the original artwork's mastery and how creative constraints force artistic innovation without relying on color or resolution.
#2
520 points · milkglass · comments
Discussion Summary
The discussion revolves around whether Western decline in manufacturing and coding skills stems from AI adoption or deeper management failures—primarily short-term cost-cutting that eliminates junior hiring and mentorship pipelines, causing irreplaceable tacit knowledge to vanish. Commenters debate whether this is an inevitable consequence of progress and specialization, or a preventable crisis rooted in prioritizing quarterly profits over long-term institutional capability.
#3
464 points · skullone · comments
Discussion Summary
Trump's firing of all 24 National Science Board members has sparked concern among scientists and commenters about damage to U.S. research independence and global competitiveness, with debate over whether this represents ideological purging, misguided policy, or necessary disruption.
#4
439 points · pr337h4m · comments
Discussion Summary
An amateur mathematician used ChatGPT to solve a 60-year-old Erdős problem by having the AI apply a known formula from a related field in an unexpected way, sparking debate about whether this represents genuine AI intelligence, the role of serendipity versus capability, and the practical value of AI in mathematical discovery.
#5
357 points · gwerbret · comments
Discussion Summary
The discussion highlights that while the USB cheat sheet is useful, commenters criticize USB's confusing naming conventions and suggest technical corrections about wire counts, SBU definitions, and missing specifications like USB Low Speed and Thunderbolt 5.
#6
326 points · robinhouston · comments
Discussion Summary
The discussion centers on a 2012 project that created 3D-printable adapter pieces to make different construction toy systems (Lego, K'nex, Tinkertoys, etc.) compatible with each other, with commenters appreciating the interoperability concept while noting practical limitations and the corporate lock-in that exists in the toy industry.

AI / Machine Learning

243 points · chirau · comments
Discussion Summary
The public dislikes AI primarily due to job displacement fears, environmental costs, unethical training practices, and the industry's tone-deaf messaging prioritizing corporate profit over societal benefit, though some argue the backlash is overstated and ignores AI's legitimate applications in medicine and other fields.
149 points · Murfalo · comments
Discussion Summary
The community is skeptical of OpenAI's $25k biosafety bounty, criticizing it as underfunded marketing theater with gatekeeping restrictions, NDAs that prevent publication, vague requirements, and a winner-take-all payout structure that disincentivizes participation compared to previous open bounties.
279 points · speckx · comments
Discussion Summary
Many developers are successfully reviving abandoned personal projects using AI coding tools, finding them particularly effective at eliminating re-entry friction and handling the tedious parts, though some critics worry this reduces learning and meaningful engagement with the work itself.
137 points · marvinborner · comments
Discussion Summary
A new lambda calculus benchmark reveals that top AI models remain neck-and-neck while smaller models lag significantly, though critics argue single-attempt testing doesn't properly account for LLM non-determinism and that iterative approaches would provide more meaningful results.
14 points · jryio · comments
Discussion Summary
No comments available.

Startups / Business

11 points · xmas123 · comments
Discussion Summary
The commenter argues that ROS already provides a standard for robot APIs, but manufacturers resist adopting it; they also contend that REST's request-response model is poorly suited for robotics, which benefit more from pub/sub or blackboard architectures.

More Stories (38)

265 points · Ariarule · comments
124 points · larelli · comments
119 points · srean · comments
156 points · thehappyfellow · comments
76 points · offbyone42 · comments
123 points · signa11 · comments
32 points · frereubu · comments
27 points · aaronbrethorst · comments
Tell HN: An app is silently installing itself on my iPhone every day
311 points · _-x-_ · comments
Ask HN: How are you using AI code assistants on large messy legacy code bases?
13 points · thinkingtoilet · comments
38 points · Cider9986 · comments
12 points · mooreds · comments
Ask HN: How many tabs do you have open in the browser(s) and why?
11 points · juujian · comments
13 points · j_m_b · comments
Created by Zipper Data Co.  · 2026-04-26 12:00 UTC  · Unsubscribe

Get digests like this delivered to your inbox every morning.

Subscribe Free