DeepSeek-v3.2: Pushing the frontier of open large language models [pdf]

Digital Spaceport is a really good channel, I second that - the author is not sparing any detail. The cheaper options always use CPU only, or sharding between different cheap GPUs (without SLI/switching) - which is not good for all use cases (he also highlights this). But some his prices are one-off bargains for used stuff. And RAM prices doubled this year, so you won't buy 2x256 GB DDR4 for $336, no matter what: https://digitalspaceport.com/500-deepseek-r1-671b-local-ai-s...

Dunno, DeepSeek on MacStudio doesn't feel much slower than when using it directly on deepseek.com; 6t/s is still around 24 characters per second which is faster than many people could read. I also have 6000 Pro but you won't fit any large model there and to be able to run DeepSeek R1/3.1/3.2 671B at Q4 you'd need 5-6 of them depending on the communication overhead. MacStudio is the simplest solution to run it locally.

The hardware can always contain undetectable sub-devices that can magically intercept anything with no possibility for the software to detect this.

You should remember that all iPhones had for several years an undetected hardware backdoor, until a couple of years ago, when independent researchers have found it and reported the Apple bugs as CVEs, so Apple was forced to fix the vulnerabilities.

The hardware backdoor consisted in the fact that writing some magic values to some supposedly unused addresses allowed the bypassing of all memory protections. The backdoor is likely to have consisted in some memory test registers, which are used during manufacturing, but which should be disabled before shipping the phone to customers, which Apple had not done.

This hardware backdoor, coupled with some bugs in a few Apple system libraries, allowed the knowledgeable attackers to send remotely an invisible message to the iPhone, which was able to take complete control over the iPhone, allowing the attacker to read any file and to record from cameras and microphones. A reboot of the iPhone removed the remote control, but then the attacker would immediately send another invisible message, regaining control.

There was no way to detect that the iPhone was remotely controlled. The backdoor was discovered only externally in the firewalls of a company, because the iPhones generated a suspiciously high amount of Internet traffic, without apparent causes.

This has been widely reported at the time and discussed on HN, but some people continue to be not aware about how little you can trust even major companies like Apple to deliver the right hardware.

The identity of the attackers who exploited this Apple hardware backdoor has not been revealed, but it is likely that they had needed the cooperation of Apple insiders, at least for access to secret Apple documentation, if not for intentionally ensuring that the hardware backdoor remained open.

Thus the fact that Apple publishes only incomplete technical documentation has helped only the attackers, allowing them to remain undiscovered for many years, against the interests of the Apple customers. Had the specifications of the test registers been public, someone would have quickly discovered that they had remained unprotected after production.

Therefore, for many years the iPhones of certain valuable targets had magically intercepted all their communications and they have sent them to an unknown country (due to the nature of some of the identified targets and the amount of resources required to carry the attacks, it has been speculated that the country could have been Israel, but no public evidence exists; a US TLA is the main plausible alternative, as some targets were Russians).

I'm not seeing anywhere in that page anything about an assumed static basket of human wants and needs. Maybe I missed it -- can you point out where you saw that?

Interesting, though, that per the very same article someone like Adam Smith concurred empirically with Marx's observation on the titular tendency of rates of profit to fall. This suggests to me it likely had some meat to it.

I used DeepSeek-v3.2 to solve two coding problems by pasting code and directions as one large prompt into a chat interface and it performed very well. VERY WELL!

I am still happy to pay Google because of their ecosystem or Gemini app, NotebookLM, Colab, gemini-cli, etc. Google’s moat for me is all the tooling and engineering around the models.

That said, my one year Google AI subscription ends in four months and I might try an alternative, or at least evaluate options. Alibaba Cloud looks like an interesting low cost alternative to AWS for building systems. I am now a retired ‘gentleman scientist’ now and my personal research is inexpensive no matter who I pay for inference compute, but it is fun to spend a small amount of time evaluating alternatives even though mostly using Google is time efficient.