This is a big bet. Look at what happened in 2001 with the dot-com boom. We're still trading on their dark fiber over-build today. Meanwhile any overcapacity built in fabs will quickly be made obsolete by newer and better fab technology (or at least, that's been the pattern for the past 30 years).

I think you're missing the fact that building new supply takes time and sustained commitment, and there's simply nobody in a good position to make that commitment without losing big if your thesis turns out to be wrong.

If the demand for new AI builds is eventually satisfied, or worse, craters overnight, then who will be left holding the bag? It sure won't be Google or Apple, or even NVIDIA - it will be TSMC and Samsung.

But as you mention, this is all temporary. Either you're right, and demand will remain sustained long enough for some of the providers to decide to take that risk, or demand will crater and prices will fall.

The fact that the S&P 500 is near record highs at the same time as consumer confidence is at a 70 year low is not encouraging for continued all steam ahead in my mind... but then it's easy to predict a general future recession, and much harder to predict it to the day.

I doubt consumers will regret these compute purchases in 3 years, my 3 year old GPU is still holding value, actually increased in value, and I use it more now than ever.

I did predict or expect that if oil went to $150 we'd be in recession territory, currently hovering below that level and folks are feeling the squeeze and aren't happy. Things could get worse or better, tough to tell, but I think silicon demand is more or less secular and will do relatively well in a variety of macro conditions.

The fact that we _haven't_ seen such deals be made, and that ~50% of new datacenter builds have been quietly cancelled[1], suggests to me that we're dealing with paper demand more than true sustained demand.

1: https://www.tomshardware.com/tech-industry/artificial-intell...

Bad metaphor. Not being forward enough is a common mistake. When learning, you may want to make yourself _feel_ like you're too far forward. We can commonly think we're too far forward when actually overall we're still too far back. Teaching it is very hard.

Although it is dependent on your style of skis (e.g. carving) and style of skiing (powder, racing, cruising).

For me, it is very rare that I go too far forward (I began with a leant back style and I haven't rectified that after many years of skiing). I try to prioritise fun over ability.

If someone points out their wrongdoing, rather than take it seriously and work to purge their organization of corruption, they maliciously attack the people who pointed it out.

What did we expect when they dropped "don't be evil" from their company values?

Probably wasn't clear enough if you don't know what that is already, apologies

It's an Asus Ascent GX10, which is a little mini PC with 128GB of LPDDR5X as shared memory for an Nvidia GB10 "Blackwell" (kind of, it's a long story) GPU and a MediaTek ARM CPU

could you tell me the long story?

edit: or wait, is it quasi-Blackwell the way all DGX Sparks are quasi-Blackwell? like the actual silicon is different but it's sorta Blackwell-shaped?

The promise of this chip was “write your code locally, then deploy to the same architecture in the data centre!”

Which is nonsense, because the GB10 is better described as “Hopper with Blackwell characteristics” IMO.

Still great hardware, especially for the price and learning. But we are only just starting to get the kernels written to take advantage of it, and mma.sync is sad compared to tcgen05

    #!/bin/sh
    export ANTHROPIC_BASE_URL=https://api.deepseek.com/anthropic
    export ANTHROPIC_AUTH_TOKEN=sk-secret
    export ANTHROPIC_MODEL=deepseek-v4-flash
    export CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC=1
    exec claude $@

This is what I’ve been using for non-confidential projects for about a week now (soon after v4 came out). I honestly can’t tell the difference, but I’m not doing anything crazy with it either.

Worth noting that I don’t think DeepSeek‘s API lets you opt out of training. Once this is up on other providers though… (OpenRouter is just proxying to DeepSeek atm)

Also, you can use HuggingFace Inference for DeepSeek V4 or Kimi K2.6, both of which work quite well and route through providers that you can enable/disable (like Together AI, DeepInfra, etc) - you'll have to check their policies but I think most of those commercial inference providers claim to not train on your data either.

Why is that?

IIRC, USA data protection protects data of US citizens only, foreigners data is not protected, and the companies are not even allowed to disclose when they collect those data.

HN is an American site. If you look at the US government, it is going to fearmonger about anything China related, because they haven't had a genuine competitor for decades and they're scared and lashing out. Most US news just parrot the government line, sometimes more so than state TV would, and so it reflects here.

I also feel comfortable saying that many Americans don't care one bit what happens to foreigners, be it by action of their government or companies.

This is true. There are also many of us who do care.

This brings to mind something I heard recently about the so-called "Rule of 10". There will always be 3 people who support you, 3 people who are against you, and 4 people who have no idea what's going on and don't care.

Don't just focus on the 3 people who are being negative.

People can have problems with America and I'm fine with that. But pretending China isn't subsidizing industry (land, education, transportation) in a predatory fashion is silly. Too many companies have gone out of business because of it. We can all have our friends in China without pretending the CCP is playing the ballgame fairly. The government doesn't need to point it out. That doesn't even get into influence operations (which are especially easy on platforms like this.)

Seriously - there may be a day in the future where Western nations and China get along but it really can't/won't happen while it's holding all the industry and trying to take the Services income as well.

But tell me more.

That doesn't mean China should not be criticized. But to me it's clear that the China blame game is not about a genuine concern for Chinese people or its neighbors, it's about trying to keep it down because China should never dared to rise in the first place.

Anglo Saxons and maybe the French should be in charge and the rest should be resource colonies. It very much feels like that Western mentality is still there.

Agreed, the US definitely needs to do some introspection to sort out its own shit (and stop spraying it on everyone else).

However, that does not mean that China gets a pass. Fundamentally, the Chinese model of governance does not protect the individual. For all its faults, the US model is based upon the idea of individual liberty, which acts as a touchstone and allows it to self-correct whenever it goes to far in the wrong direction. That's something the Chinese model does not do, and means that, short of a revolution, it will continue to be an authoritarian state with all of the malignant features that entails.

Look, am not here to defend the Chinese model but I find it interesting how convinced you seem that individualism is the right model for everyone.

While I would generally agree with you, I have spoken to many from poorer countries who say that they prefer to trade some individualism for a steady hand of economic development and lifting the population from poverty. That is the Chinese model.

These people would argue that they can reclaim more and more individual freedom as the country gets richer and more self confident.

I am not saying they are right, but looking at a nominal democracy like India and a nominal autocracy like China, I know which government works better as far as raising the living standards of its population and it's not the Indian one.

My hope is that China will continue to liberalize on its own. Forcing it will likely only reverse the gains.

Individualism also leads to the sort of healthcare system the US had or Skid Row. So it's not all roses.

What's the point of this kind of statement for you? Does this help you understand others or just continue to drive the wedge in? Where are you from? Ask yourself can the statement,

"many {of my country} don't care one bit what happens to foreigners, be it by action of the government or companies" not be read as true?

There are self-absorbed, disinterested, uncompassionate people in every country which will satisfy your "many" qualifier.

However looking at the polls in the US gives you a fairly decent idea that there's a decent chunk of people that seem to get off on violence towards non-Americans. Why do you think ICE went with the violent tactics it did?

As to

> What's the point of this kind of statement for you? Does this help you understand others or just continue to drive the wedge in?

The point is to maybe make some Americans ask what it is that they can do to reform the government they have the most direct influence over (their own) instead of trying to reassure themselves that theirs is still better than country's X.

I set ZDR to true, and it only calls from the third party ZDR Deepseek APIs. Bit more expensive, but my client wants it.

----

alias clauded='ANTHROPIC_BASE_URL=https://api.deepseek.com/anthropic ANTHROPIC_AUTH_TOKEN=$DEEPSEEK_API_KEY ANTHROPIC_MODEL=deepseek-v4-pro[1m] ANTHROPIC_DEFAULT_OPUS_MODEL=deepseek-v4-pro[1m] ANTHROPIC_DEFAULT_SONNET_MODEL=deepseek-v4-pro[1m] ANTHROPIC_DEFAULT_HAIKU_MODEL=deepseek-v4-flash CLAUDE_CODE_SUBAGENT_MODEL=deepseek-v4-flash CLAUDE_CODE_EFFORT_LEVEL=max claude'

----

I doubt that the opus, sonnet, and haiku model args actually matter if you want to omit them.

I run this on a VPS that has no other credentials or project access so I can give it the skip permissions arg.

Also, the author checked in their apparently effective social media advertising plan: https://github.com/aattaran/deepclaude/commit/a90a399682defc... (which seems to be working)

All I'm trying to say that generalizing like suggested might exclude some useful things.

    deepseek() {
      unset ANTHROPIC_AUTH_TOKEN
      local -x ANTHROPIC_BASE_URL="https://api.deepseek.com/anthropic"
      local -x ANTHROPIC_AUTH_TOKEN="${DEEPSEEK_API_KEY}"
      local -x ANTHROPIC_MODEL="deepseek-v4-pro"
      local -x ANTHROPIC_SMALL_FAST_MODEL="deepseek-v4-flash"
      local -x API_TIMEOUT_MS=600000
      local -x CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC=1
      TMUX= command claude "$@"
    }

is flash version on level of gpt 5.4 mini

PS. mentioning amp because i used to use it and I pay directly for token. I topped up 5 usd so I will be going to use it and see how far can it take me. But my impression so far is even when model subsidization is done, those open source models are quite viable alternatives.

My understanding is that DeepSeek V4 Pro is going to be uniquely good at working on consumer platforms with SSD offload, due to its extremely lean KV cache. Even if you only have a slow consumer platform, you should be able to just let it grind on a huge batch of tasks in parallel entirely unattended, and wake up later to a finished job.

AIUI, people are even experimenting with offloading the KV cache itself to storage, which may unlock this batching capability even beyond physical RAM limits as contexts grow. (This used to be considered a bad idea with bulky KV caches, due to concerns about wearout and performance, but the much leaner KV cache of DeepSeek V4 changes the picture quite radically.)

AI currently works basically by sending your entire codebase and workflow, and internal communication over the internet to some third party provider, and your only protection is some legal document say they pinky promise they won't train on your data.

And said promise is made by people whose entire business model relies on being able to slurp up all the licensed content on the internet and ignore said licensing, on the defense of being too big to fail.

> AIUI, people are even experimenting with offloading the KV cache itself to storage, which may unlock this batching capability even beyond physical RAM limits as contexts grow.

Especially this point. Any reason that this idea was considered bad? Is it due to the speed difference between the GPU VRAM to the RAM?

> Any reason that this idea was considered bad?

Because the KV cache was too big, even for a small context. This is still an issue with open models other than DeepSeek V4, though to a somewhat smaller extent than used to be the case. But the tiny KV of DeepSeek V4 is genuinely new.

Model inference was never subsidized. Inference is highly profitable with today's prices. That's why you have many inference providers. My guess, the prices for inference will go down, as more competition starts cutting the margin.

It's model training, development and R&D that cost a lot, and companies creating closed models don't have any business model except astroturfing and trying to recover training costs through overpriced inference.

The token cost makes it tempting to use for token-heavy tasks like this

I have been wanting to try CC with different models since Opus went downhill last month..

What limitations or issues have you noticed when using DeepSeek with Claude Code if any?

$ ollama launch claude --model qwen3.5 [1]

[1] https://docs.ollama.com/integrations/claude-code

I'm currently shopping for a local LLM setup and between something like the Framework Desktop with 64-128GB of shared RAM or just adding a 3090 or 4090 to my homelab so I'm very curious what hardware is working well for others.

Parameters are like Hertz - they don't really tell you much until you know the rest anyways. In this case, a parameter is a bfloat16 (2 bytes). I'm sure someone will bother to makes quants at some point.

> I'm currently shopping for a local LLM setup and between something like the Framework Desktop with 64-128GB of shared RAM or just adding a 3090 or 4090 to my homelab so I'm very curious what hardware is working well for others.

I grabbed a 395 laptop w/ 128 GB to be a personal travel workstation. Great for that purpose. Not exactly a speed demon with LLMs but it can load large ones (which run even slower as a result) and that wasn't really my intent. I've found GPUs make more usable local LLMs, particularly in the speed department, but I suppose that depends more on how you really use them and how much you're willing to pay to have enough total VRAM.

It's next to impossible to make your money back on local (regardless what you buy) so I'd just say "go for whatever amount of best you're willing to put money down for" and enjoy it.

So a wyde

A kid and I sometimes engage in a game where they try to get me to look where necessary, so far without success.

  > FaceID only unlocks if you actually look at the screen.