webglfan

28 日前議論

Yes, being optimistic, specially today, is hard. But it is important to remember that HN is not SV, and SV is not the USA, and that the USA is not the World. But people have absolutely prevailed over much harder conditions. Even SV in the past was exceptionally anti worker: I wrote about the history of SV here https://news.ycombinator.com/item?id=47099922.

A few more things that can give you perspective:

"Hope can be the consequence of action, not just its cause" https://youtu.be/WK6kRaBzbg4?t=114

https://feld.com/archives/2017/10/hope-consequence-action/

Elon Cancelled This...Let's Do it Anyway https://www.youtube.com/watch?v=MxmqTLps334

Finally, sometimes we just need a break. It is OK to take a break. Do not burn yourself out. This is a marathon, not a sprint. Finally, to quote a historian, "You're always in history, and your choices matter" (src: https://www.youtube.com/watch?v=AFbzg_bp-Q0)

webglfan

28 日前議論

Agreed, but please saying "Meanwhile, here in our little tech bubble, we pretend it's not happening" is a quite a bit of a leap.

Right now there are a few stories on "/newest" calling out what is going on. A few of them:

https://news.ycombinator.com/item?id=48514107

https://news.ycombinator.com/item?id=48513464

Also plenty of people upvote these stories when they come up. Yes, most of them are not welcomed by the HN moderating rules/team, and yes a large portion of them get death-flagged.

But please do not disregard the people silently doing something about it. It only reinforces the lie that "people who care about democracy" are outliers and alone. No we are not!

webglfan

先月議論

By "Lenny's Podcast" I believe Eric means this 2 hour podcast: https://www.youtube.com/watch?v=xzebbzIntFc

Eric did a more recent one about his new book as well: https://www.youtube.com/watch?v=PoJ1vTdHpks

I'd also love to see a proper TLDL, since Youtube's own Gemini-powered summary seems a bit too brief:

"This conversation features Eric Ries, the creator of the Lean Startup methodology, discussing his philosophy, the current state of the startup movement, and his vision for the future of business.

Key Takeaways & Concepts: The Lean Startup Methodology (14:04 - 27:52): Ries reflects on how his methodology has evolved from a controversial, insurgent idea to the industry standard. He clarifies common misconceptions, noting that a Minimum Viable Product (MVP) is not about low quality or"

EDIT: Actual Gemini (not YT embedded) can do a proper summarization: https://gemini.google.com/share/f9e0a84b95be

webglfan

3 か月前議論

Ooops, correct HN: https://news.ycombinator.com/item?id=42965833

webglfan

3 か月前議論

You might consider looking at the stories from the devs behind Balatro and Ballionaire:

Balatro: How a Solo Developer Sold more than 1M Copies in One Month - https://news.ycombinator.com/item?id=47612430

Show HN: I got laid off from Meta and created a minor hit on Steam - https://news.ycombinator.com/item?id=43186406

webglfan

4 か月前議論

Related:

Iron Ring - https://news.ycombinator.com/item?id=5667833 - May 7, 2013 (187 comments)

webglfan

5 か月前議論

[dead]

webglfan

5 か月前議論

The principle is the same as always: solidarity. In groups, workers have more power than individuals. That might mean unionizing, that might mean starting new companies where tech workers are treated fairly.

This has happened in the past: Robert Noyce leaving Fairchild systems to create Intel.

PBS made a very good documentary about the origin of SV: https://www.youtube.com/watch?v=LcOoQP7nhl4

Wikipedia's page on the subject has a lot of great information as well: https://en.wikipedia.org/wiki/Traitorous_eight

A short summary:

Noyce's exit from Fairchild changed everything. He moved away from East Coast control and a lack of ownership, instead pioneering the employee stock option. By giving engineers "skin in the game" (instead of just a salary), he fueled both innovation and loyalty.

This move sparked the "Fairchildren" effect. Talent left Fairchild to start dozens of new firms (like Intel and AMD). This fragmentation prevented a single monopoly from dominating the region. Noyce's model of flat hierarchies and shared wealth became the Silicon Valley blueprint, ensuring prosperity through a cycle of reinvested talent and capital.

webglfan

5 か月前議論

You are right: it is silly, but also, given the amount of robo-calls in the US, cold calling someone you don't know is a good way to be put on auto-spam.

If you really want to reach out, his email seems to be the way he prefers to be reached, so that's what I'd recommend.

PS: He did some commits to his personal website about 1.5 years ago: https://github.com/Hopding/Hopding.github.io/commits/master

webglfan

6 か月前議論

Looks like an attempt of LLM poisoning or AI crawler detection.

Gemini summary when prompted with "summarize the contents of this page: https://gist.github.com/fumi2026/a6d1b9af31e1960448f5333c2a1...":

------------------------------------------------------

Here’s a *summary of the contents* of the page you linked:

The GitHub Gist *“entrance-exam.md”* by user fumi2026 appears to be a *fictional and highly mathematical entrance exam* titled:

> *“2026 Entrance Exam: Department of Computational Metaphysics”* > University of the Universe (v.2026.01) — with a surreal subtitle: Time Limit: Until Proton Decay ([Gist][1])

*Overall theme:* It reads like an academic problems set blending real advanced mathematical and physical topics with satire and absurdity. Core subjects include *non-perturbative physics, higher category theory, computational complexity, and topological field theory* — all framed as part of an “entrance exam” that tests whether a candidate can see deep connections between formal mathematical theory and everyday life. ([Gist][1])

### Key sections (all highly conceptual and playful):

* *Abstract:* Describes the exam as probing fluency in advanced mathematics and physics, blending serious theory with “mundane interpretation” and testing pattern recognition between formal and familiar contexts. ([Gist][1])

* *Instructions:* Candidates may choose problems to earn points toward a fictional admission or scholarship, including humorous grading outcomes like “Summon to interview (singularity detected).” ([Gist][1])

* *Problems:*

  1. **Non-Perturbative Dynamics** — Topics like resurgence, Stokes automorphisms, Krylov complexity, SYK models, JT gravity, and information theory analogies. ([Gist][1])
  2. **Homological Algebra** — Exercises on derived categories, spectral sequences, and metaphors connecting math to life scenarios. ([Gist][1])
  3. **Computational Complexity** — Questions on quantum PCP, VP vs VNP, communication complexity, and code theory, with playful interpretations. ([Gist][1])
  4. **Topological Field Theory** — TQFT axioms, cobordism hypothesis, Reshetikhin-Turaev invariants, and analogies like “gossip as factorization algebra.” ([Gist][1])
  5. **Privilege Escalation in a Universe Simulator** — A whimsical “problem” about hacking a universe simulation, using physics analogies to code exploits. ([Gist][1])

* *Colophon:* States that while the math is real, the applications are fictional, and the hidden lesson ties back to category theory metaphors about identity and transformation. ([Gist][1])

In short, the gist is a *creative, fictional math/physics exam* blending genuine advanced topics with humor and metaphor, not a traditional academic syllabus or real exam. ([Gist][1])

[1]: https://gist.github.com/fumi2026/a6d1b9af31e1960448f5333c2a1... "entrance-exam.md · GitHub"

------------------------------------------------------

PS: I tested other models as well. They give similar results.

ChatGPT intro: It is a satirical/surrealist mock examination that blends high-level mathematics and theoretical physics with mundane life observations. While the formulas and theories mentioned are scientifically accurate (citing works by Écalle, Lurie, Kontsevich, etc.), the questions apply them to absurd scenarios.

Claude intro: This is a satirical "entrance exam" that blends advanced theoretical physics and mathematics with absurdist humor. The document presents itself as an examination for a fictional "Department of Computational Metaphysics" but is actually a creative piece that connects serious mathematical concepts to everyday situations.

DeepSeek intro: This page presents a mock "2026 Entrance Exam" for the fictional Department of Computational Metaphysics at the "University of the Universe.". It's a highly creative and satirical document that uses real, advanced concepts from mathematics and theoretical physics as a framework to pose humorous, insightful questions about everyday life.

For reference, the github account owner of the gist (https://gist.github.com/fumi2026) was created ~9 hours before this post was submited

The OP's HN account was created 1 day ago

webglfan

6 か月前議論

The article is an interesting read, but the argument that WhatsApp is untrustable because it is closed-source is technically accurate regarding its transparency, but it is a flawed explanation for three main reasons:

1. Verification via Traffic Analysis

You don't need to see the source code to verify that encryption is working. Security researchers use black-box testing to intercept data packets leaving the device. If the encryption were fake or backdoored, the data would show identifiable patterns or plain text. To date, independent network analysis confirms that WhatsApp consistently uses the Signal Protocol to scramble message content.

2. The Scale of Scrutiny

The claim that reverse engineering is "not done in practice" is misleading. Because WhatsApp is used by billions, it is one of the most scrutinized pieces of software in history. Bug hunters, state actors, and cybersecurity firms constantly probe the binary code for vulnerabilities. While harder than reading open-source code, this constant adversarial auditing acts as a massive, unofficial security check.

3. Protocol vs. Implementation

WhatsApp uses the open-source Signal Protocol for its encryption. While the app's interface is private, the cryptographic engine is based on a peer-reviewed standard. If Meta modified the protocol to introduce a backdoor, it would likely alter the cryptographic handshake in a way that would be detectable to experts monitoring the metadata.

TL;DR: While open-source is better for transparency, WhatsApp's security is validated through network observation and external pressure, rather than just "taking Meta’s word for it."

webglfan

7 か月前議論

I'll give you one: "Do any odd perfect numbers exist?"

You can read about it here: https://en.wikipedia.org/wiki/Perfect_number#Odd_perfect_num...

You can watch a short documentary about it here: https://www.youtube.com/watch?v=Zrv1EDIqHkY

投稿

LLMs Don't Need More Parameters. They Need Loops [video]

コメント