Hacker Newsnew | past | comments | ask | show | jobs | submit | qphe95's commentslogin

If a company had a model that could cure cancer they would be incentivized to release the cure ASAP before they get decapitation striked by regulators and other AI "safetyists".


We are literally in the golden age of privacy with AI generated private apps, so I wouldn't worry too much about this.


You guys should really be making an end to end PCB machine, sourcing the parts and part inventory for custom PCBs really should be the responsibility of the end user.


Possibly, but our users find value in not having to kit, store, or ship parts to CMs.


I'm skeptical this is a cost effective for low volume prototyping which is where the speed of made in America matters.


whats the closest to this that exists today? (??? + pick and place?)


I'm not the expert but Gemini doesn't say anything exists besides the Voltera V-One (which obviously doesn't work for complex PCBs). I am imagining something like a PCB microfactory the size of a container (or even smaller) that completely handles PCB etch and drill, solder paste application, PnP, baking, and testing.


That would be super cool. Difficulty comes with multi-layer PCBs, hard to find a solution for laminating.


PCBA lines are largely automated as of today, issues start when you do high-mix work. Most time ends up being spent switching between jobs and programming machines, meaning that tool usage time is low.


Theres no actual evidence they didn't just distill Kimi K3


At this point, it doesn't matter who is distilling from who.


Is there any actual evidence that they did?


There's also no evidence they didn't just distill Gemma 3. And also no evidence its not a purple popsicle.


Can't wait to distill Deepseek v4 flash to America-1


Well banning it pushed it to Radicle and now people are moving to GitHub alternatives so I think the Streisand effect worked.


Did banning it cause it to get adopted even faster in India than it otherwise would have? That seems like the salient point to me for the Indian government’s calculus. I don’t have a particularly good feel for Indian society but my prior is that this kind of official friction puts a pretty low hard cap on how popular an app can get. If the realistic ambition was only ever to reach a small circle of activists and nerds then sure, this event probably helped the app reach those people faster.


People can be surprisingly resourceful when they feel like they have a need to do so.

In tech circles, we talk a lot about friction and user experience, and those matter enormously, but they're not the be-all-end-all. We have OSS developers underestimating friction on one hand, and tech companies overestimating it on the other.

Even if something requires you to be tech-savvy to set up, people will still do it if the need is dire enough, because everybody "knows a guy". Console modchips in the 2000s are my go-to example. They were devilishly complex to set up, often requiring proficiency with a soldering iron, with stakes for failure being very high (breaking the console you just spend all your savings buying), yet everybody and their mom had one. All you needed was a friend with such a chip, who could tell you who the "guy" in your town is.


I dont have data for any country, but if anything banning it from github gave it visibility (I personally learnt of it because of that). I don't think that banning it from github hurts its distribution really because I do not expect github to be the channel it gets popularised through as opposed to social media, in which case where the source code is hosted does not really matter. If it gets banned from the app stores, though, it will be a different issue that will definitely hurt.


I never even heard of the app until I saw the news it was getting banned, so I installed it out of curiosity - and I'm keeping it installed because I like the idea of it. I presume I'm not alone in this. They're basically giving it free publicity by attempting to ban it.


We both know systems are more complex than # of downloads. In this case, the ban was a sign of weakness the establishment of the government and Github and definitely their control was weakened, which is the opposite of their goal of maintaining control, so it did not have the desired effect.


Likely faster than it would've otherwise but not likely as fast as it needs to to succeed. Similar to when Twitter got bought out and then Musk banned people who linked their Mastodon. It certainly helped Mastodon in the short term but likely hurt it in the long term


As soon as installation of software outside Play Store is made impossible, any such discussion is moot already anyway.


It pushed it to a place with no search-ability and significantly less users. It worked to effectively suppress the project. I say that as someone who uses Radicle and likes the project (I am hoping node search comes soon).


Are those things that have happened?

I mean, here's the first 4 URLs that show up when I Google for bitchat:

  https://bitchat.free/
  https://play.google.com/store/apps/details?id=com.bitchat.droid&hl=en_US
  https://github.com/permissionlesstech/bitchat
  https://apps.apple.com/us/app/bitchat-mesh/id6748219622
All of those work for me where I am, so AFAICT: It hasn't been pushed from one place to another at all. It's still where it was.

It has additionally been copied over to Radicle... and that's OK, isn't it?


Another day of AI another day of AI not being able to maintain useful features people used to take for granted


I'm reading the code for this and I'm like yeah it doesn't scale. Like forget about debugging this code, does anybody even know what it is doing?


This is how you end up with gstack. When you believe too hard in secret metrics because when people know about metrics it stops working you inflict psychosis chasing correlations that don't mean anything.


People said AI was going to replace all programmers, yet AI can't even maintain a x64 macOS port of Java.


LLM code isn't allowed in OpenJdk.


Practically that means code that is identifiable as LLM, if the code is indistinguishable it is indistinguishable.


Isn't this the paid version of the JDK?

EDIT: I should actually click the link, not just read the title. Rip.


The paid version of the JDK is OpenJdk.


Thats an arbitrary decision that can easily be reversed.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: