HN comments - Digest ⚙️ Edit Settings

Period: 2025-01-19 00:00 - 2025-01-26 00:00 📚 All Digests

AI Digest

新评论摘要

1. 关于“谢谢HN:我的自主创业公司今天被收购了”

2. 关于“展示HN:我从零开始制作了一款开源笔记本电脑”

3. 关于“斯塔格项目:软银、OpenAI、Oracle、MGX建设数据中心”

4. 关于“罗斯·乌尔布里希特被完全赦免”

5. 关于“0-click 去匿名化攻击 Signal、Discord 和其他平台”

6. 关于“创业冬天:黑客新闻失去了信心”

7. 关于“我见过保罗·格雷厄姆一次”

8. 关于“DeepSeek-R1”

9. 关于“我再也不会使用 GitHub Actions”

10. 关于“Tiktok表示恢复美国用户服务”



Details

bestcomments

  • New comment by firefoxd in "Thank HN: My bootstrapped startup got acquired today"
  • Content:

    My employer picked vwo over my own custom in house ab testing tool [0]. I was so pissed!

    Congrats, yours was the better tool.

    [0]: https://news.ycombinator.com/item?id=23471050


  • New comment by t4TLLLSZ185x in "Show HN: I made an open-source laptop from scratch"
  • Content:

    For those that missed it, this Engineer is in _high school_.

    Byran, I have been a professional engineer longer than you have been alive, I can tell you right now that I have met very, very few people that would have the motivation, skill and sticktoitivness to pull this off.


  • New comment by nrp in "Show HN: I made an open-source laptop from scratch"
  • Content:

    Super impressive, and awesome to see that you were able to use Framework Laptop hinges. Let me know if you need more. We have a ton of remaining 3.3kg ones!


  • New comment by mppm in "Stargate Project: SoftBank, OpenAI, Oracle, MGX to build data centers"
  • Content:

    Apart from my general queasiness about the whole AGI scaling business and the power concentration that comes with it, these are the exact four people/entities that I would not want to be at the tip of said power concentration.


  • New comment by wolfgang42 in "Ross Ulbricht granted a full pardon"
  • Content:

    Tangentially related: I had the disconcerting experience of reading a Wired article about his arrest[1] while unknowingly sitting about six feet from the spot where he was apprehended. When I read that the FBI agents had stopped at Bello Coffee while preparing their stakeout, I thought, huh, interesting coincidence, I just had a coffee there.

    Then Ulbricht walked into the public library and sat down at the table directly in front of me, and suddenly as I was reading I could look up and see exactly the chair he had been in, where the plainclothes police had positioned themselves, how they had arranged a distraction.

    Having this tableau unexpectedly unfold right in front of my eyes was a fascinating experience, and it certainly made the article suddenly get a lot more immersive!

    [1] https://www.wired.com/2015/05/silk-road-2/

    EDIT: to be clear, I was not present for the arrest. I was reading the magazine, some years after the arrest, but in the same place as the arrest. (I didn’t qualify the events with “I read that...” since I thought the narrative ellipsis would be obvious from context; evidently not.)


  • New comment by thrwaway1985882 in "0-click deanonymization attack targeting Signal, Discord, other platforms"
  • Content:

    "Near a user" is also a big assumption. I'm ~200 miles to ORD and ~500 to IAD, but my ISP's peering & upstream arrangements mean Cloudflare serves my traffic 700 miles from DFW.

    But, at the same time: Cloudflare isn't going to serve me a cache from Seattle, Manchester, or Tokyo. Pinning down an unknown Signal user to even a rough geographic location is an important bit of metadata that could combine to unmask an individual. Neat attack!


  • New comment by internet_points in "0-click deanonymization attack targeting Signal, Discord, other platforms"
  • Content:

    So if you send a picture to a Signal user, it's retrieved via cloudflare, and cached in a data center near that user; now you can look up the cache status and find the data center used. I'd say "deanonymization" is stretching it, unless the user is in the middle of nowhere (no other users near the data center). But interesting writeup anyway.


  • New comment by KaiserPro in "Startup Winter: Hacker News Lost Its Faith"
  • Content:

    I think the bigger issue is that most of the financial upsides of startups were illusory

    Unless you're a founder, you're unlikely to make any money, and predatory VC/backers using convertibles/senior debt to turn your equity into Zimbabwean dollars. Thats before any stats are calculated for lasting long enough to either IPO, make money or get bought out.

    As someone who did successfully exit, the stories of multiple-millions being handed out turned out to be legends. Dodgy accounting & staff retention meant that actually you don't get fuckyou money.

    The only realy reason to join a startup is that you either believe in the "mission" or that you believe it will bring you joy to work in that way.


  • New comment by afavour in "I Met Paul Graham Once"
  • Content:

    In the long run I think realizations like the authors are healthy ones.

    PG is not a hero. He's just a guy. A guy who entered into business transactions with a number of people, many of whom benefitted greatly (as did Paul himself). I'm not saying any of that as a negative! Just that we have a habit of attributing superhuman characteristics to folks (Obama getting the Nobel Peace Prize comes to mind) and ending up disappointed.

    I'm not an affected group by any means but I still share the disappointment in the world we see today vs the possibilities I felt tech would allow when I was younger. The tech CEOs I previously viewed as visionaries now just look like a new generation of socially regressive robber barons. I wanted to be one of those CEOs, these days I'm still not quite sure what I want to be. My only consolation is knowing that I'm seeing the world more accurately than I once did.


  • New comment by simonw in "DeepSeek-R1"
  • Content:

    OK, these are a LOT of fun to play with. I've been trying out a quantized version of the Llama 3 one from here: https://huggingface.co/unsloth/DeepSeek-R1-Distill-Llama-8B-...

    The one I'm running is the 8.54GB file. I'm using Ollama like this:

        ollama run hf.co/unsloth/DeepSeek-R1-Distill-Llama-8B-GGUF:Q8_0
    
    You can prompt it directly there, but I'm using my LLM tool and the llm-ollama plugin to run and log prompts against it. Once Ollama has loaded the model (from the above command) you can try those with uvx like this:

        uvx --with llm-ollama \
          llm -m 'hf.co/unsloth/DeepSeek-R1-Distill-Llama-8B-GGUF:Q8_0' \
          'a joke about a pelican and a walrus who run a tea room together'
    
    Here's what I got - the joke itself is rubbish but the "thinking" section is fascinating: https://gist.github.com/simonw/f505ce733a435c8fc8fdf3448e381...

    I also set an alias for the model like this:

        llm aliases set r1l 'hf.co/unsloth/DeepSeek-R1-Distill-Llama-8B-GGUF:Q8_0' 
    
    Now I can run "llm -m r1l" (for R1 Llama) instead.

    I wrote up my experiments so far on my blog: https://simonwillison.net/2025/Jan/20/deepseek-r1/


  • New comment by arghwhat in "I'll think twice before using GitHub Actions again"
  • Content:

    > no way of running actions locally

    My policy is to never let pipeline DSLs contain any actual logic outside orchestration for the task, relying solely on one-liner build or test commands. If the task is more complicated than a one-liner, make a script for it in the repo to make it a one-liner. Doesn't matter if it's GitHub Actions, Jenkins, Azure DevOps (which has super cursed yaml), etc.

    This in turn means that you can do what the pipeline does with a one-liner too, whether manually, from a vscode launch command, a git hook, etc.

    This same approach can fix the mess of path-specific validation too - write a regular script (shell, python, JS, whatever you fancy) that checks what has changed and calls the appropriate validation script. The GitHub action is only used to run the script on PR and to prepare the CI container for whatever the script needs, and the same pipeline will always run.


  • New comment by Buttons840 in "TikTok says it is restoring service for U.S. users"
  • Content:

    The TikTok debate has always been about the balance between national security and free speech.

    We found a compromise. TikTok will remain, all of its national security risks will remain. Also, the law that tramples free speech is upheld by the court, but will be blantently ignored and unenforced.

    Everybody loses. This outcome is worse than anyone could have conceived.


  • New comment by Brystephor in "TikTok says it is restoring service for U.S. users"
  • Content:

    What's that saying? The best way to get a promotion is to cause a problem and then fix it?

    Political things aside, it's crazy to see so much of a flip-flop so quickly. Has there been any other behavior like this in the past where a company "shut themselves down" to make a big political statement and then almost immediately undid the shut down?


  • New comment by robswc in "TikTok says it is restoring service for U.S. users"
  • Content:

    The straight up "shout out" in the pop-up, I almost couldn't believe my eyes.

    I don't think I've seen anything like it in a long time. I also don't think an American company would ever do that as it seems "unprofessional." Ironically, it probably got them huge bonus points so they know what they're doing.


  • New comment by daemoens in "TikTok goes dark in the US"
  • Content:

    The app was shutdown a couple of hours ago in the US and this was the message all TikTok users saw when they opened the app.[1]

    The same guy who pushed for a ban massively last year, is going to save the app despite the security concerns he and most of our government said they had. If only we knew what happened in that classified briefing that made them vote together across party lines.

    [1] https://a57.foxnews.com/static.foxbusiness.com/foxbusiness.c...


  • New comment by wilg in "TikTok goes dark in the US"
  • Content:

    None of those arguments are the salient one, which is that a geopolitical adversary has control over a major influence vector on US public opinion. They could simply have divested.