Hacker News Digest

Friday, May 15, 2026

In This Issue

  • Hacker News
  • Removing the Modem and GPS from My 2024 RAV4 Hybrid
  • RTX 5090 and M4 MacBook Air: Can It Game?
  • MIT: 20% drop in incoming graduate students
  • New arXiv policy: 1-year ban for hallucinated references
  • God Damn AI is making me dumb
  • Mullvad exit IPs are surprisingly identifying
  • What happens when you post a real Monet and say it's AI?
  • Grok Build
  • Overseas fakers using AI videos to push a narrative of UK decline, BBC finds

Zipper Data Brief

May 15, 2026
Your daily digest of the best from Hacker News

Top 6 Trending

#1
916 points · arkadiyt · comments
Discussion Summary
A 2024 RAV4 owner removed the car's modem and GPS to prevent data collection, sparking debate about whether this actually works—commenters note that Bluetooth tethering to your phone and CarPlay/Android Auto may still transmit telemetry, though these claims lack concrete evidence, and warn that future cars may require connectivity for security features or risk being bricked.
#2
615 points · allenleee · comments
Discussion Summary
The discussion praises an impressive technical achievement of getting an RTX 5090 eGPU working with an M4 MacBook Air for gaming and local AI inference, with commenters highlighting the practical benefits for LLM prompt processing speed while also expressing frustration with Apple's lack of official eGPU support and native game compatibility.
#3
605 points · dmayo · comments
Discussion Summary
The 20% drop in MIT graduate students is primarily driven by federal funding cuts and the new 8% endowment tax on wealthy universities, though commenters debate secondary factors like international student visa concerns, AI hype reducing motivation, and broader systemic issues with academia's low pay and exploitative conditions.
#4
533 points · gjuggler · comments
Discussion Summary
Most commenters support arXiv's ban on hallucinated references as necessary to combat low-effort AI-generated submissions that waste reviewers' time, though some raise practical concerns about enforcement and whether the policy definition is clear enough.
#5
497 points · Eighth · comments
Discussion Summary
Developers are split on whether AI coding tools make them smarter or dumber: experienced programmers using AI strategically for routine tasks while handling complex decisions feel more capable, while those relying heavily on AI without deep engagement worry they're losing critical thinking skills and struggling to learn effectively.
#6
428 points · RGBCube · comments
Discussion Summary
Mullvad's exit IPs are deterministically assigned based on user keys, allowing correlation across multiple servers and potentially identifying the same user—a privacy flaw the company acknowledges and is patching. However, commenters debate whether this matters given VPNs aren't designed for anonymity like Tor, and most users just need privacy from ISPs rather than complete anonymity.

AI / Machine Learning

83 points · nailer · comments
Discussion Summary
The discussion reveals that people's judgment of art is heavily influenced by context and attribution rather than the work itself—when a real Monet was labeled as AI, viewers critiqued it negatively, exposing how much expectation shapes perception. However, commenters debate whether this proves anything meaningful about AI art, with many arguing the experiment is cherry-picked and that art's value fundamentally depends on human creation and intent, not just technical quality.
90 points · meetpateltech · comments
Discussion Summary
The discussion expresses skepticism about Grok Build's value proposition, with commenters questioning whether it offers meaningful advantages over existing alternatives while criticizing its high subscription costs ($300/month for full features) and lack of clear differentiation beyond potential Twitter integration.
32 points · dijksterhuis · comments
Discussion Summary
The discussion debates whether overseas AI-generated propaganda about UK decline is more concerning than the underlying issues it exploits, with commenters noting this fits a broader pattern of algorithmically-amplified divisive content across platforms that monetizes fear and outrage.
358 points · mikeevans · comments
Discussion Summary
Users are excited that Codex can now be accessed remotely from mobile devices, enabling "vibe coding" workflows where developers can iterate with AI agents on-the-go and have drafts ready for review on desktop. However, there are mixed experiences reported—some find mobile coding less effective due to smaller screens and lack of keyboard, while others appreciate the ability to manage active work from anywhere without being tethered to their desk.
30 points · skp1995 · comments
Discussion Summary
I don't have access to browse external links or view the HackerNews discussion you're referencing. To help you summarize it, could you please paste the comments or provide the key points from the discussion?

Startups / Business

120 points · Lihh27 · comments
Discussion Summary
The discussion celebrates Germany's €1.3M funding for KDE as recognition that software is a public good, with commenters praising KDE's efficiency and impact on the open-source ecosystem, while one skeptic questions the investment's viability without a profit model.
10 points · ent101 · comments
Discussion Summary
A developer notes that Z.ai offers free integration but monetizes through user limits, raising questions about mainstream adoption given this pricing model's departure from typical practices.
Ask HN: What's the hardest part of building a SaaS that users keep paying for?
10 points · specwiseai · comments
Discussion Summary
The hardest part of building a sustainable SaaS is creating something genuinely useful that users can't live without, which requires understanding your target customers deeply, starting with free usage to learn their pain points, and building habits around your core feature rather than just acquiring users.
121 points · surprisetalk · comments
Discussion Summary
Commenters express skepticism about the partnership, questioning whether it's genuine charitable work or a financial arrangement benefiting both parties, while criticizing the Gates Foundation's track record and raising concerns about the optics of billionaire-backed AI initiatives.
48 points · knuaym9 · comments
Discussion Summary
Most commenters praise the route-sharing concept as innovative and trustworthy, but identify critical challenges: weak cold start (need curated content, not UGC initially), limited geographic availability (US-only frustrates international users), and low user retention since people don't travel frequently.

More Stories (34)

342 points · caust1c · comments
83 points · robin_reala · comments
Claude Account Suspended Seconds After Purchase?
70 points · AnthropicWHAT · comments
55 points · maxwellito · comments
200 points · rbanffy · comments
19 points · steven_noble · comments
83 points · tambourine_man · comments
Which country will be the first to pass laws limiting Meta Ray-Ban glasses?
23 points · nothrowaways · comments
17 points · devitoria · comments
16 points · ulrikrasmussen · comments
110 points · Einenlum · comments
43 points · yminsky · comments
197 points · jsploit · comments
14 points · roywashere · comments
Created by Zipper Data Co.  · 2026-05-15 12:01 UTC  · Unsubscribe

Get digests like this delivered to your inbox every morning.

Subscribe Free