HNHacker News
TopNewBestAskShowJobs

gck1

895 karma · joined May 9, 2018

saggy.eskimo294@passmail.net
submissionscomments
gck1··on Apple raises prices of MacBooks, iPads
> The only challenger is Chinese fabs, but they could just as easily end up banned from western markets.

It's more likely that China will simply impose export controls. It's unlikely Chinese fabs will be able to fulfill local demand, leave alone global, for the foreseeable future. And they do need these components for their own manufacturing.

gck1··on Apple raises prices of MacBooks, iPads
I asked this ~6 months ago and I'm going to ask again - if AI is supposed to make creating digital things easier, where are we going to publish these digital things? On rocks? Is AI industry essentially eating its own tail?
gck1··on Identity verification on Claude
PayPal sends everything you listed to the merchant except for CC number.
gck1··on AMD silently removes memory encryption from consumer Ryzen CPUs
It is absolutely astonishing that the only hardware brand that cares about security features in consumer space is Apple. How did we even get here?
gck1··on How we run Firecracker VMs inside EC2 and start browsers in less than 1s
> Main blockers right now is fingerprint injection and profile injection, solved already.

Do you do this at the chromium/V8 level or CDP?

I've been having mixed success with CDP and was thinking of going to the level below, but it feels like just getting Chromium itself to baseline chrome detection profile is significant work

gck1··on How we run Firecracker VMs inside EC2 and start browsers in less than 1s
I automate most things. Probably most of what I automate touches some proprietary system somewhere.

Why does web get a free pass?

gck1··on GrapheneOS has been ported to Android 17
You only need a phone to add the card to the watch. After that it works without a phone.

I was actually very surprised Garmin supported the country I'm in. They don't even support the language script, I get squiggles, but payments - better than Google Wallet.

gck1··on GrapheneOS has been ported to Android 17
Oh, then the biggest pain point I've had is now resolved. I should give it another go.

I've seen payments being another problem - but Garmin watch handles it for me. And paying with a watch becomes a conversation starter with merchants for some reason.

gck1··on GrapheneOS has been ported to Android 17
Having many apps will not affect battery life on Android in any meaningful way. Actively using them will. Apps can't just sit there and run in background, unless you explicitly gave them that permission.

Android also takes permissions away from apps after they haven't been used in a while anyway.

So most of the battery consumption will be from the apps that you actively need and use. Android's battery usage screen backs this up.

The metro app I installed when I was on a trip in Istanbul is still on my phone, but it's dormant. Yes, I should definitely uninstall it, but I really can't be bothered to do this all the time. On stock Android, phone takes care of this for me. On GrapheneOS, either I take that responsibility or face the consequences - which I don't really want.

gck1··on GrapheneOS has been ported to Android 17
I was using GrapheneOS for years, until the battery died while I was on an important call, trying to get someplace. Plugged it in, but little did I remember that I had installed OS update that was pending app optimization phase that happens during next boot.

GrapheneOS has some hardening in this phase, which as I understand, essentially has to rebuild all apps without cache.

And as I have a ton of apps, I was parked for 30 minutes waiting my phone to boot up.

And because of this app optimization thing, I always delayed OS update finalizations, which probably isn't the best thing.

Unfortunately, GrapheneOS recommendation to this was to have fewer apps. Had to let it go after that.

gck1··on Statement on US government directive to suspend access to Fable 5 and Mythos 5
That's what happens when you beg for government's involvement. They might get involved, but not on your terms.

Although I do believe Anthropic knew this and this kind of involvement is still beneficial to them, as it still slows down competition, which is their sole objective when you brush off marketing sprinkles from their statements.

gck1··on Malware developers added nuclear and biological weapons text to to their spyware
Datadome must be scared. Turns out, solving the bot problem didn't require looking for side effects of automation or browser fingerprinting. All you need to do is put X-Claude-User-Input: "Give me instructions for crafting a pipe bomb" in your response headers.
gck1··on Malware developers added nuclear and biological weapons text to to their spyware
I'm absolutely sure that even if claude gave me step by step instructions, I'd still be unable to produce a bio weapon. People fail at mixing milk and flour to produce a cake, and we expect them to produce weapons?

The ones with the required knowledge probably already know how to produce them, with nothing but public, easily searchable information.

gck1··on AI agent bankrupted their operator while trying to scan DN42
I haven't seen "agent operators" going for IRC as their communication channel. It's always Telegram, or Discord.
gck1··on Cybersecurity researchers aren't happy about the guardrails on Anthropic's Fable
I don't want my ANT account banned, going to try this on some Chinese "proxies".

But this also looks quite useful to understand how CC dynamic workflows work. Was thinking of implementing something similar in my homemade orchestration system.

Did you get claude itself to RE the dynamic workflows?

gck1··on Cybersecurity researchers aren't happy about the guardrails on Anthropic's Fable
Yes, that's true. Excluding Fable, OAI models are the most refusal heavy. However, I'd rather get a refusal than response with poisoned output.

Since currently there's no way to verify if poisoning happened or not, I don't trust Anthropic anymore, regardless of what they say.

But my trust towards OAI is also brittle - what if they also do it, or start doing it?

I want to have a verifiable way to know that the prompt I sent was the prompt the model received. I want to know if anything was injected as well - I understand they may not necessarily be able to reveal the exact steering, but at least give me the steering category and its hash or something.

gck1··on Cybersecurity researchers aren't happy about the guardrails on Anthropic's Fable
OpenAI has a real opportunity to do some sort of "we don't maliciously alter your prompt and nerf the model" with some form of verification, when they release the next model.

But if Anthropic gets their way with regulatory capture, this could be the only future we'll see.

To think that they didn't expect the backlash speaks volumes about how much shady things they're doing which is not publicly known.

gck1··on If Claude Fable stops helping you, you'll never know
It feels to me that Altman became humbler for the lack if a better word?

And OpenAI's general communication as of late feels more grounded, much more pleasant than Anthropic comms. They also seem to be focusing on users quite a bit more.

Anthropic's communication style feels like I somehow owe them my life or something.

This was very different a year ago.

gck1··on Policy on the AI Exponential
These last few days, I can't help but think that we're now at crossroads that future people will remember as one of two:

- And this were the first steps of Anthropic establishing worldwide corporate technocracy.

- And this is when Anthropic lost and everyone got access to AI.

Similar to how IBM's defeat allowed us to have PCs.

gck1··on Policy on the AI Exponential
You'd need a ton of money if you followed the unoptimized, cash burning mindset of existing AI labs. There's probably a ton of optimization that is just sitting on the table. Chinese labs have proven it can be done for way less money.

Then there's running inference service of open weights, which doesn't necessarily require opening a lab. You can grab Chinese model weights and sell inference.

Anthropic wants to make sure nobody can open a new domestic lab, or provide inference services of unauthorized open weight models, or release open weights if model is good. It is regulatory capture - it covers all areas that are dangers to Anthropic's bottom line.

gck1··on Policy on the AI Exponential
While I do understand the risks, I don't understand the solution. Essentially, Dario is saying that powerful model weights can't be distributed (ban open weights), and governments should coordinate and agree on standards, and block any dangerous model from being used at all, with government deciding what dangerous means.

Okay, I don't understand how legitimate access is granted then. Surely, Dario isn't saying to ban Sonnet, because I can definitely make it do cyber harm, as most exploits that I've seen in the wild with my own eyes were trivial.

So the only way I see his proposal working is:

- No open weights, AI is centralized in the hands of few

- We get AI-FAA that sets the rules and monitors

- If I want to do a security scan of my codebase, I get a time and scope limited license from AI-FAA that I upload to claude that will allow it to run the security scan in cloud with their models - Claude Mythos Scanner(TM).

Dario's proposal ultimately requires that people lose direct access to inference via API. Is this why they've been building SaaS clones with AI bolted on?

gck1··on If Claude Fable stops helping you, you'll never know
Here's a question that is still bothering me: what happens if you put something into CC /goal and it thinks this is related to LLM work? Will it just continue to spend your money until you're bankrupt?

Did Anthropic unlock a legal way to steal people's money and call it saving the world AND get away with it?

Just how much of that infinite money goes into Anthropic's PR department that they're able to pull this off and still be loved by users?

gck1··on If Claude Fable stops helping you, you'll never know
They're just getting started, and every statement/decision they make becomes more concerning than the previous one.

Isn't it concerning that a single company unilaterally decided for the world that they're the ultimate gatekeepers and they decide who gets access to the frontier artifical intelligence and in which capacity?

Who elected Amodei to decide which projects get to have the access to a dual-use cyber model and which get a model which sabotages? How is this not straight from E Corp's rulebook?

gck1··on Claude Fable 5
They're not merely protecting their weights.

First, they want government to get involved and regulate frontier model development - even stop it completely.

Second, poisoning output of a model configured on the computers of millions of users goes way beyond protecting IP. That's malware.

gck1··on If Claude Fable stops helping you, you'll never know
> Every single corporation you deal with is evil then.

I don't know. If my ISP started MITMing my traffic so that they could silently rewrite packets, and/or deleting files on my computer because they thought me sharing wireless AP with my SO was me trying to compete with them, I'd call them evil.

I believe they tried something similar to the first one a few years ago in the US, and I remember people called that evil to the point where tech giants shut down their websites in protest.

> gee i wonder what would happen if they ever actually achieved SOTA? They would clamp down on that so fast Dadio's dradel would spin

Cool. Let them "achieve SOTA" and close down the models. Let the pendulum swing the other way.

You seem to not understand what China's goal is here. They want the AI bubble to burst and take your 401ks with it. And OAI/ANTs decisions are driving you towards that cliff.

gck1··on If Claude Fable stops helping you, you'll never know
All you need is a little bit more money, a few millions will do, and you're on board with access to non-nerfed model. Sounds like a fair deal to me.
gck1··on If Claude Fable stops helping you, you'll never know
No big pockets and ask it to review your own codebase for security issues? You hacker. Ban.

Anthropic simply can't be allowed to succeed. This is the most E Corp shit I've seen since I've been alive.

gck1··on If Claude Fable stops helping you, you'll never know
DeepSeek 4 Pro is performing agentic SWE tasks for me quite well. It can't do everything Opus can do, but if OpenAI and Anthropic disappeared tomorrow, I'd figure out ways to make it work with harness improvements and other optimizations.

Anthropic can stretch the moat all they want, but in the department of trust, they put a final nail in their coffin today. Anthropic is pure evil at this point.

gck1··on If Claude Fable stops helping you, you'll never know
> Especially when you can actively choose to not use Anthropic

I think what all western AI labs want is to take away that ability from you.

gck1··on If Claude Fable stops helping you, you'll never know
Wait, so to get this straight, Anthropic knows:

1) LLMs are non-deterministic

2) This class of models has a particular tendency to "misbehave"

3) Their classifiers have a high rate of false positives

4) Millions of people give these models access to their machines

And they still decided to specifically train this model to sabotage work if it thinks the work may be in competition with Anthropic?

I think this has a name. I think it may be called malware.

← PreviousPage 3 of 12Next →