Hacker News Digest
Saturday, April 4, 2026
In This Issue
- Hacker News
- Artemis II crew take “spectacular” image of Earth
- Show HN: I built a frontpage for personal blogs
- US F-15E jet confirmed shot down over Iran as Tehran releases wreckage images
- Oracle Files H-1B Visa Petitions Amid Mass Layoffs
- iNaturalist
- A School District Tried to Help Train Waymos to Stop for School Buses
- Claude 4.6 Jailbroken
- Lower Price for ChatGPT Business
- The Hardest Document Extraction Problem in Insurance
Zipper Data Brief
April 04, 2026
Your daily digest of the best from Hacker News
Top 6 Trending
#1
804 points
· andsoitis
· comments
# Summary
The Artemis II crew captured a stunning photograph of Earth's nightside illuminated by moonlight using a Nikon D5 camera, sparking discussion about the image quality, technical details, and its profound reminder of Earth's fragility and beauty from space.
#2
Tell HN: Anthropic no longer allowing Claude Code subscriptions to use OpenClaw
778 points
· firloop
· comments
# Summary
Anthropic is restricting Claude Code subscriptions from using OpenClaw due to capacity constraints and to protect service quality for typical users, though the move is controversial—some view it as a reasonable infrastructure decision while others see it as unfairly penalizing power users who are paying for the service.
#3
718 points
· ramkarthikk
· comments
# Summary
The HN community enthusiastically praised this blog aggregator as a nostalgic return to web discovery and curated directories, comparing it favorably to existing indie web tools, though some noted practical issues like pagination bugs and infinite scroll blocking the footer, plus broader concerns about discoverability and content curation.
#4
509 points
· tjwds
· comments
# Summary
The discussion centers on a U.S. F-15E shot down over Iran, with commenters debating the strategic implications (noting Iran's air defenses weren't fully destroyed as claimed), expressing concern for the downed pilots, and raising broader questions about the justification and consequences of U.S./Israeli military operations in Iran.
#5
490 points
· kklisura
· comments
# Summary
The discussion is largely skeptical of the headline's framing, with many commenters arguing it's misleading because the H-1B filings occurred in different fiscal years than the recent layoffs, and noting that Oracle laid off workers globally—not just Americans. Others debate whether the $100k visa fee makes H-1B hiring economically rational, and discuss broader questions about whether H-1B programs genuinely fill talent gaps or depress American tech worker wages.
#6
448 points
· bookofjoe
· comments
# Summary
The discussion highlights iNaturalist as a valuable citizen science platform with an excellent API and friendly community for wildlife documentation, while raising concerns about privacy risks from publicly exposed user locations and the organization's shifting priorities away from power users toward growth and new app development.
AI / Machine Learning
32 points
· phlummox
· comments
# Summary
Commenters debate whether Waymo's school bus stopping incidents represent serious safety failures or are being overblown—some note improvement trends and that human drivers also fail to stop, while others argue self-driving vehicles must be held to higher standards and demand enforcement since Waymo repeatedly violates traffic laws around schools.
22 points
· NuClide
· comments
# Summary
The discussion is skeptical about claims of a Claude "jailbreak," with commenters questioning whether this represents actual vulnerability versus expected LLM behavior; one commenter notes a more serious finding about extracting 915 files from Claude's sandbox including tokens and IPs, while others debate the semantics of "jailbreak" versus "uncensored" and reference a researcher known for regularly testing frontier models.
16 points
· alxthm
· comments
OpenAI is lowering ChatGPT Business pricing to shift users toward pay-per-token consumption models, while some businesses want more flexible seat-based pricing options for their needs.
31 points
· sgondala_ycapp
· comments
The discussion highlights a self-correcting document extraction system that achieved 95% accuracy in insurance processing, with broader insights on how AI agents can be made reliable through self-correction loops, success criteria validation, and continuous improvement frameworks.
74 points
· borski
· comments
Most commenters praise the travel hacking toolkit as a clean, useful tool with low barriers to entry, though some question whether optimizing points redemption remains worthwhile in 2026 given the shift toward straightforward cash-back strategies. Several users highlight limitations like incomplete coverage of niche hotel booking sites and credit card-specific travel benefits that the tool should address.
Startups / Business
17 points
· abawany
· comments
# Summary
Critics view Musk's requirement as a potential grift tactic—forcing IPO banks to buy Grok subscriptions appears to artificially inflate xAI's value while raising capital for SpaceX, raising concerns about the IPO's legitimacy.
348 points
· carabiner
· comments
# Summary
Delve, a compliance-as-a-service company, was removed from Y Combinator after allegations of serious fraud emerged, including selling fake SOC2 compliance reports to customers (potentially exposing them to legal liability under regulations like HIPAA) and repurposing open-source software without proper attribution. The removal highlights a fundamental breach of trust within YC's community model, particularly damaging given that Delve's core business was supposed to be ensuring compliance.
9 points
· spenvo
· comments
No comments available.
15 points
· rallies
· comments
# Summary
A startup is running an experiment where LLMs manage real money in the stock market with access to financial data and tools, showing mixed results since November 2024. Commenters raise skepticism about whether this differentiates from existing quant funds, question the methodology (including concerns about selective display of negative returns), and debate whether AI-driven trading is genuinely novel.
13 points
· jcbhmr
· comments
The comments are deeply skeptical of Tan's claim, arguing that lines of code is a meaningless productivity metric prone to gaming, and that AI-generated code without proper review creates technical debt and bugs rather than genuine value.
More Stories (34)
332 points
· chrisjj
· comments
275 points
· edent
· comments
109 points
· uticus
· comments
228 points
· nickslaughter02
· comments
151 points
· janvdberg
· comments
96 points
· rguldener
· comments
44 points
· johnbarron
· comments
215 points
· lxm
· comments
45 points
· brandonb
· comments
107 points
· Fudgel
· comments
35 points
· ZunarJ5
· comments
82 points
· matthewsinclair
· comments
I prefer OG style websites – what are yours?
31 points
· gorfian_robot
· comments
400 points
· kykeonaut
· comments
199 points
· upofadown
· comments
141 points
· veqq
· comments
54 points
· indy
· comments
24 points
· signa11
· comments
23 points
· npongratz
· comments
16 points
· frenchtoast8
· comments
17 points
· remilouf
· comments
17 points
· iLemming
· comments
14 points
· kogasa240p
· comments
17 points
· uelbably
· comments
58 points
· angst
· comments
99 points
· jandeboevrie
· comments
396 points
· chalmovsky
· comments
167 points
· uticus
· comments
181 points
· 0o_MrPatrick_o0
· comments
26 points
· zdw
· comments
49 points
· dmitrygr
· comments
43 points
· cg505
· comments
43 points
· luca-ctx
· comments
36 points
· geox
· comments
Created by Zipper Data Co.
· 2026-04-04 12:01 UTC
· Unsubscribe
Get digests like this delivered to your inbox every morning.
Subscribe Free