I bet Sam secretly pledged to DoD that the red lines were only temporary, for optics and to calm employees at the all hands meeting.
A few months down the line, OpenAI will quietly decide that their next model is safe enough for autonomous weapons, and remove their safeguard layer. The mass surveillance enablement might be an indirect deal through Palantir.
The fact they were able to strike a deal in the first place hours after Anthropic was declared a supply chain risk should make this obvious. Their employees are smart people, the only way they can reason past this is their compensation.
That's very possible! In the last few days Anthropic was getting a lot of attention, and OpenAI was looking weaker in comparison. It seems like a politically coordinated job to remove competition.
For sure, he's been pissed that OpenAI no longer has the Mandate of Heaven and Claude is all anyone has been talking about since December. (And it's not just an ego thing: because OAI isn't profitable yet, they need the hype to keep going to raise money on favorable terms, so loss of buzz is an existential threat). I absolutely believe that he started making calls to try and get buddies in the White House to take Anthropic down.
I don't trust Sam to be telling the truth. It would be to his benefit to lie about this and make Anthropic look bad, so he of course would, even if it's not actually the case.
Hell, I would have thought it likely that anthropic was doing the same thing. Of course that was proven wrong, but for OAI I wouldn't even be guessing. This has always been what sama does.
It makes sense if you imagine the real motivation is “make sure the AI contracts go to my good friend Sam”, and all the red line stuff is just a way to pick a fight with Anthropic.
What about the N-600 form which the article highlights as an impossible barrier for many immigrants to attain their certificate of citizenship. That isn't hard?
Big difference! That's a full VM, while Termux is more like a Debian container. For most use cases you will have a better time with Termux, which also ships useful Android integrations such as clipboard and notifications.
They note that the robot achieved an "intermediate" skill level. This has been determined by letting it play matches against people of different skill levels as determined by a professional table tennis coach. The "Results" section explains this.
Google had a very competitive employee ping pong league and one of the coaches was on the USA Olympic team, I doubt they would lie about ping pong skills
Googlers are well known for speaking up when they disagree with something the company is doing, but yeah I mostly wanted to call out how big the ping pong culture is at Google at least in Mountain View
But they don't explain what metrics they use to differentiate skill level other than waving the hands of a puppet coach. The arm control is great but they don't show it returning non-softballs. Is that what "intermediate" means?
Probably, if you consider that a beginner can rarely volley at all. At an intermediate level, a fast return pretty much just wins the point, when it doesn't hit the net or miss the table.
GrapheneOS, in my opinion, can only exist the way they do because of their exclusive support for Pixel devices. It allows them to provide a uniquely high quality ROM with proper support guarantees.
Compare that to other custom ROMs, where you typically depend on volunteers maintaining the various devices. Sometimes people decide to step down, and suddenly find your device unsupported. This happened to me with LineageOS/CyanogenMod.
My understanding is also that the OEM ROMs of Pixel devices are closer to AOSP than those of other vendors like Samsung. This simplifies the maintenance of the ROMs, and enables the project to develop meaningful features instead.
See https://news.ycombinator.com/item?id=43670303 for a detailed explanation. If we were going to support another device, it would need to meet the security requirements AND provide proper production quality non-stock OS support with a strong commitment to keeping it properly working.
We can't safely use a device where they might patch out support for what we rely on. As an example, OnePlus patched out support for alternate OS verified boot due to serious security vulnerabilities with how they'd implemented it. Operating systems relying on it would no longer be able to update the firmware, leaving them insecure, but yet the verified boot never worked properly anyway so it ended up worse than them not trying in the first place.
Realistically, what we expect is that Pixels are the only devices we're not involved in making which we'll be able to support for the foreseeable future. To support other devices, we need a partnership with a competent OEM like Samsung. We can raise money to pay the usual licensing fees for platforms and then work with them where we have paid support so they aren't going to break things for us, drop update support unexpectedly, etc.
Thank you for working on GOS. It is one of the few, if not only, OS that makes mobile devices usable for me.
> To support other devices, we need a partnership with a competent OEM like Samsung.
How realistic would it be to have an official Graphene phone? Could you partner with an open-source friendly company like Purism or Framework to design and manufacture the hardware to your specifications? That would be the ultimate mobile device for tech and privacy nerds, for which I'm sure many would be willing to pay a premium over mass manufactured devices.
As much as I trust your vetting of Google devices, there's a strong incongruity between the mission of a trillion-dollar adtech company and yours that I just can't reconcile.
Samsung is the main company we'd really want to work with. They already provide devices meeting nearly all our requirements, they just don't provide us a way to support them yet. If they started doing that, we could support some of their devices. Better yet, if they actively worked with us, we could help them make major security improvements and there could be first class GrapheneOS support. We can raise money for it rather than having to convince them that it makes business sense to fill a product niche they aren't currently providing.
> As much as I trust your vetting of Google devices, there's a strong incongruity between the mission of a trillion-dollar adtech company and yours that I just can't reconcile.
Apple and Google provide a high level of security for their mobile devices. Other OEMs aren't on the same level. Samsung is the only one that's even remotely close. We'd love to work with Samsung but wanting to do it doesn't mean they will work with us. We could potentially pay them to build a device for us if we raised enough money in advance. We choose devices based on their security and other properties along with the actual record of the company making it. Corporations are amoral profit seeking entities in general. That's not something specific to Google.
> Could you partner with an open-source friendly company like Purism or Framework to design and manufacture the hardware to your specifications?
Framework doesn't currently build devices close to meeting our requirements but is not anti-security or misleading people like Purism. We wouldn't have any issue working with them, we just don't really expect them to start building what we need any time soon.
Purism has extremely anti-security practices and makes extraordinarily insecure devices incredibly far from meeting our requirements. They purposely choose low security components and don't provide important updates. They even block providing the updates from the OS. We'd much rather support a low-end Motorola phone with only 2-3 years of support than a Purism device because they're so much less secure than mainstream hardware. Purism has 0 days of update support for their products in the sense that we require.
We don't consider Purism to be a privacy friendly company due to the atrocious security of their products and services. The software they use on top is also far less private and secure than the Android Open Source Project. It's largely the complete opposite of GrapheneOS. They also do a lot of false marketing that's directly harmful to us such as their false claims about cellular radios. In reality, their device has a much less secure cellular radio with far more attack surface exposed from the OS than an iPhone. It's less isolated, not more isolated, and yet they've convinced many people otherwise with their marketing and done harm to the whole space with the misconception people now have. The same applies in other areas.
If we partnered with Samsung, people would know they're going to get a good product and that the company making the hardware would still be around providing support years later. If we instead partnered with a company known for not shipping people what they ordered and not providing what was claimed in the promotional material, that would be very hard for people to trust. Our community would be shocked and incredibly disappointed if we did that.
Could you protect against NetFlow analysis by pushing a bunch of noise over the VPN tunnel at all times? I'd assume it would at least make the analysis significantly more challenging.
Some of the prior works in this paper[0] address noise in anonymity networks, but in general: you either add noise at the link level which malicious nodes can identify & ignore, or you add noise by injecting fake chaff packets that are dropped somewhere inside the network which are statistically identified when you look at packet density across the network.
This might or might not extend to VPN nodes depending on your threat model - I'd personally assume every single node offered to me by a company in exchange for money is malicious if I was concerned about privacy.
