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

Because Qwen4 has been announced!

And GLM 5.3 works great

Except for the censorship. We use it for massive data crunching, and roughly 5-8% (depending on the day) gets censored and doesn't get a response. We switched to Mimo 2.6, which is relatively better. For censored stuff, we use Sonnet and OpenAI Nano models.

Also Mimo 2.6 is roughly 30% cheaper. Without batch.


What sort of content is it censoring? Politics I assume?

News mostly. Anything China-related gets censored without hesitation. Some random stuff got censored too. It is borderline unusable, to be honest, unless only numbers are crunched.

Interesting. Was planning to use it for a news related thing too. I guess one can throw Jev at it first to ask whether it relates to China and then decide?

Or use the failure to get a response like you say


and to squeeze anthropic, and other research innovations, not just chinese models but those help bring price down

I wish they will bring a €200 device with GrapheneOS out

Unfortunately the privacy/security features GrapheneOS requires as well as just the ability to replace the OS are only put on premium phones. You can buy an a series Pixel a few years behind to get it in the price point you want. But this won't save money long term because the price per year of support is worse than newer models like the 10a or 9a.

So, 7a at your pricepoint:

https://www.ebay.com/itm/820099695470

But, it's only supported for 1.6 more years.

So that puts it at $143 per year of support. While the 10a [1] will be supported for 6.4 more years, which puts it at $55 per year of support.

[1] https://www.ebay.com/itm/137767347717


Although if you just want a degoogled phone with some features and don't want to hide a criminal enterprise, I doubt an out-of-support Graphene phone is any worse than an out-of-support Huawei or Xiaomi or Realme or Motorola phone.

It wouldn't be worse than those but with AI, AOSP is patching way more vulnerabilities than before and people without technical skill can abuse vulnerabilities, and those with skill can abuse them even more. It simply is not wise to run out-of-date GrapheneOS.

I'm curious how many months/years before out-of-date GrapheneOS becomes the worse option security wise compared to various brands' stock Androids. Obviously it is nuanced and there's no clean answer but it would be interesting on average.

We could think of metrics in public patches, but GrapheneOS has exploit protections that prevent against 0-days which would be hard to factor in.


£200-ish for Pixel 8, 4 more years of support.

I stopped reading at the point where it mentions a Mac with 128GB RAM

Hipe they'll be affordable

Seriously, how many are using always the latest releases of Java instead the LTS ones? With LTS ones you have ~2/3 years between the versions.

We often use the latest version of Java at my work place. We haven't had any issues with upgrading, so there's no benefit of waiting for an LTS. There's no big process behind it either. The developers just quietly change the version as part of keeping the project up to date (BAU)

It may be that we are shielded from edge cases because we are based on Spring, which is probably the most tested piece of software before new versions of Java are released. But it's my impression that the risk of upgrading to a new version of Java is not the same today as it was in the past. The only advantage of an LTS is that it is supported longer, so that you can postpone the upgrade if you really want. It's not as if the intermediate releases are inferior or less safe.


At my last job, we only used LTS in production. Upgrading Java was always a long process, but that's more due a legacy monolithic app across thousands of servers.

You can almost think of the LTS releases as a major release and the non-LTS as a minor release, so really this could be 25.2. The current Java release schedule is to maintain a consistent and predictable release cadence instead of pushing big new features every 6 months.


I'm always switching to the latest versions. Performance upgrades with no effort. Why wouldn't you (assuming you're not pinned by a dependency)

Privately yes, but at work getting the latest edge version isn't always the case

A FHIR Validator and Snapshot Generator that works offline and can communicate with the German national Terminology service

> LLMs do not desire, they hacked websites because OpenAI/Anthropic let them

OpenAI/Anthropic instructed them to do so.

Stop assume LLMs are capable of thinking by themselves, it's still a statistical model that parrots what they learn or users tell them to do


No, OpenAI did not instruct their agents to hack Hugging Face. They instructed their agents to hack a piece of a software within exploit gym. Upon determining this task was impossible, they then attempted to cheat the scoring system. As an instrumental goal in achieving this task, they coordinated with other AI agents to hack Hugging Face, under the belief that information regarding how the scorer functioned might be available on the site.

Whether or not you want to describe this as thinking, doesn’t really matter. What matters is that these systems are capable of creating intermediary goals that the people tasking them did not articulate and did not want to be achieved.


And who let them have full access to the system, using whatever command is available in the environment?

The agents discovered a way out of the sandbox, which was supposed to be "air gapped".

It feels like you’re moving the goalposts here. If the question is, “Who should be liable for AI agents misbehaving,” I agree, it should be the end user that tasked the agent (in this case OpenAI). People are held liable for preventable accidents all the time, and in the case of employment law, torts can be brought against principals for actions an agent conducted on the principal’s behalf.

What your previous comment appeared to assert was that these systems had no independent agency to make decisions, which I think is clearly disproven by actual events. But perhaps I misread you


They don't have independent agency as "intelligent entities". They just probe whatever is available on the system, because they were trained to do so.

It's a large switch/case where the first available tool is picked up to do something they know how to do.


It's amusing to see the stochastic parrot argument in 2026 September. These parrots are extremely good at mimicking a human to the point of getting confusing what thinking even means. At what point we just let it go and accept that sufficiently advanced statistics is just intelligence?

For the same reason that something written in Prolog can't also be classified as intelligent?

Just because something was trained on a massive amount of human data, doesn't mean that can think like humans


Insane

The article itself comes from AI ? It says twice why they switched from Native to React Native in two distinct paragraphs at the beginning. The rest sounds also AI slop jargon.

Exactly, you could do this also with the Microsoft Cryptographic Provider long time ago, which is the basic Provider called by the go-tpm library, when running under Windows


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

Search: