HNHacker News
TopNewBestAskShowJobs

azinman2

16,377 karma · joined June 6, 2012

If you wish to contact me, leave your email in a reply.
submissionscomments
azinman2··on Another researcher says OpenAI trained on conversations, then claimed breakthrou
https://www.apple.com/privacy/docs/Audio_Intelligence_Privac...
azinman2··on Another researcher says OpenAI trained on conversations, then claimed breakthrou
What are you talking about?
azinman2··on Rivian's gambit for full autonomy
I really like the way they look. Taste is subjective and. It absolute. Some people even like the cybertruck.
azinman2··on AI models ran real businesses: They sent $12,431 in fake invoices, lost $3,200
Fair point
azinman2··on AI models ran real businesses: They sent $12,431 in fake invoices, lost $3,200
What about OpenAI’s breach of huggingface?
azinman2··on AI models ran real businesses: They sent $12,431 in fake invoices, lost $3,200
What makes it sound fake?
azinman2··on GPT-6 Astra on OpenRouter
I’m confused. I thought it was being only rolled out to select partners for now?
azinman2··on IBM Bob
Is it that diff from code as a service? It’s human intent transcribed into something useful.
azinman2··on IBM Bob
So good
azinman2··on Top Pentagon Official Contracted Personal Lawyer to Handle Minerals Deal
And doing so risks societal fractions against the idea of lawfare. It’s a lose lose situation.
azinman2··on K2 Horizon: A connected fleet of six open models
They don’t release all the code.
azinman2··on Playa Phone
That's a terrible use of burning man and against the ethos of it. It makes me sad to know that even worked.
azinman2··on Did Meta's Big Settlement Actually Help It?
That’s a bit like blaming parents for their kids smoking cigarettes. Meta intentionally ignored warnings that bad things were happening as a result of its products, and doubled down on the addictive design.
azinman2··on DeepSeek V4 Pro 0813
Which works better for you?
azinman2··on Stealing Reasoning Traces from Proprietary LLM APIs
The reasoning blocks are not stolen/mined from the internet at large directly. They’re the result of a lot of research, time, money, and expertise into creating a reasoning model. To me the answer is quite clearly no, especially when the encrypted blocks demonstrate they want to protect it.
azinman2··on Apple Silicon and macOS VMs: 11–16× Faster LLM Inference with Llama.cpp
I don’t understand what Apple 1-9 are. At first I thought it was M series chips but there is no M9 (yet)
azinman2··on H3-metal – Native MiniMax-H3 inference for Apple Silicon
Being a world class talent is independent of financial situation.
azinman2··on Muse Glimmer: 30B-parameter model optimized for always-on local agent workflows
What about Inkling? It's a quite large model that for some reason isn't discussed much.
azinman2··on DeepSeek V4 Flash 0731
Didn’t deepseek recently announce prices will go up significantly?

Right now the US dominates everyone else in actual chips in data centers. So even if deepseek etc tries to undercut, they’re very capacity limited.

azinman2··on Ten advances in mathematics and theoretical computer science
I also don’t understand why UBI is desirable. Putting everyone on welfare means everyone is poor. This won’t end well, and it’s certainly not the case that 99% of the population will let a tiny number of people remove their income in favor of pennies for all. Political violence will come first, easily.
azinman2··on How Do We Stop Vibe Coding?
Sure, but that doesn’t mean you understand the details of implementation.
azinman2··on China’s open-weights AI strategy is winning
First, it’s more than breaking a terms of service. That language sounds very minimizing, as if the scale and effect is tiny. It’s basically espionage out in the open that bootstraps other models years into the competition, which is effectively stealing billions worth of development, compute, and leadership status.

Next, just look at the lawsuit against anthropic by the book authors. What’s the recourse for those same authors in China?

Last, one continues a leadership position for your own country vs handing over leadership to a major competitor who is not allied. This has ripple effects on the economy and ability to recruit/keep talent, which then affects social fabric and political ramifications.

azinman2··on “We have information that Moonshot distilled Fable for the development of K3”
Answered in a sister thread.

Are you from Europe or South America? Or just asking on their behalf?

They are allies of the US, which includes economic spheres of influence. The premise of the petrodollar is that allies do better by being a part of it than not, which has been true for many, many decades now. China’s rise is providing an alternative for the first time in 70+ years, but so far it’s very unclear if these benefits will truly extend to allies of China or not. For example, China sends their own laborers when building infrastructure in Africa. Sure there’s new infrastructure but also debt to CCP without any benefits of knowledge transfer or local employment.

azinman2··on “We have information that Moonshot distilled Fable for the development of K3”
For a number of reasons.

First, whether we like it or not (generally not), a fuck ton of money has been invested in US AI companies, data centers, RLHF datasets amongst other datasets, etc. If that were to go to 0 that’d be quite disastrous. Alternatively, if it goes well, it’s great for the US (and to a lesser extent allies) economy and global standing.

Relatedly, tech has been a huge power house for the US economy for decades now. If the main driver of growth goes to China, what replaces it? Along with all the potential tax money, foreign investment, etc?

Next, patriotism / nationalism. This is very much a zero-sum game that defines who owns the future. Would you rather your country win or lose this? Lose this and you start losing talent, money, global standing, etc. That furthers a cascading effect that’s very negative. It also will likely create social strife with the knock on effects leading to even more bad populist ideas that just further diminish the country and tear society’s fabric apart.

None of this is hilarious.

azinman2··on How do we stop vibe coding?
But when you lead there’s a level of trust in the people who are going to implement. You may still have a high level view of what’s going on, but that’s different from sweating the details. It also is different from bottoms up building those details where consequences become clear in thought.

I’ve noticed the pressure to keep velocity and move forward means I’m less sure of those details compared to before. LLMs are also inconsistent in the work they do - some of it is brilliant, other parts are idiotic. It’s totally different than what you’d get when leading humans.

azinman2··on Flux 3
Such as?
azinman2··on “We have information that Moonshot distilled Fable for the development of K3”
Except it’s not just a dump of the internet, which Moonshot also did themselves (and probably used even more pirated content as laws in China are different without any recourse for the entire world). I don’t know why this is so unclear to folks.
azinman2··on “We have information that Moonshot distilled Fable for the development of K3”
Unless you’re from mainland China, you should.
azinman2··on China’s open-weights AI strategy is winning
I assume you don’t use any LLMs then
azinman2··on China’s open-weights AI strategy is winning
The Chinese models poked and prodded better models for training data and to avoid having to pay humans to RLHF themselves. You can call it infringement, theft, whatever, but it’s quite obviously “not ethical” to me.
← PreviousPage 2 of 34Next →