russli1993

I won't blame Altman here. Memory companies secretly love this. LLM will always be commercialized after people discovered it if it's not Altman it would be another person. And you cannot genuinely say people do not find LLM useful. You cannot genuinely say people will not use LLMs hence there is no demand for compute capacity. OpenAI and others are simply satisfying the compute demand here. Memory companies have secretly loved for an application to drive up the prices of memory chips. And you can't blame them, these foundry workers, physics Phds working at fab companies have worked much harder than your software web app engineers while earning far less. But what they build is a critical enabler of everything digital and much higher barrier to entry. If hyperscaler don't want to pay for exorbitant high memory prices, they can invest and build their own. But why don't they? This is exactly what you are describing free market economy, pricing is based on scarcity. Maybe its time for the world to actually value hardware and semiconductor for their worth.

like anyone don't have a 401k, if AI crashes, say goodbye to 401k. This is not really constructive to the situation. What is constructive is thinking about how to improve the country's electricity system, which will benefit the people and industry alike. This country charges generally 10 to 40% of taxes yet industry have to buy their own generators for power. You know a lot of capex spending by companies which is fueling the AI bubble debate is to cover job functions that should be provided by government with existing tax dollars. And these politicians have no problem voting for 1+ trillion defense spending a year.

I feel the same. I feel machine learning is 1) you need strong math knowledge but also intuition. 2) You need actual production level data, a lot of the nuances of machine learning, applied scientists' roles are how to build strong models with real world, messy, inconvenient data. You need to apply various techniques and not just memorize a specific formula.

Compared to software coding, you build websites, apps etc. The barrier to entry is much higher.

I am trying to pivot into machine learning myself, so far my understanding is try to find a software engineer role where you deal with data, model infra, data pipeline, etc. Then within the same company, try to do more modeling work and build intuition and resume credentials in this area. Someone correct me if I am wrong.

could push kv-cache to off chip cheaper storage like dram. Inference engine manages decode for a single user to run to completion before decode a different users so kv-cache fetch from dram to sram is limited.

what about tiling? seems better to call into existing kernels, fused megakernels than rewrite kernel in java.

since openAI and Anthropic are public about their anti-distillation mechanism, its either their anti-distillation doesn't work or kimi got here because their own training and RLs. Which one is it?

more to say US strategy to strangle Chinese economy and technology industry through semiconductor embargo will fail.

these labs have access to far better NVIDIA chips than deepseek

last year was google winning AI, this year is meta winning AI. I think what is true is Meta, Google will all find their place in the AI world and continue to make huge amount of money. A consumer will use many different AI products, each with a different use case. Everyone will make enough money to live well.

Huawei is launching a mobile soc built with logic folding. They take a single logic ip, like a cpu, and spread it's components across to dies and use hybrid bonding to combine top to bottom. For example the CPU SRAM on top and logic gates directly below. This reduces critical path, increases fmax, reduces power consumption. They claim this method allow them build faster, more power efficient, more function units with duv process than traditional duv. Will be interesting to benchmark after launch later this year

CSPs making own hardware still needs hardware companies, they reduce the Nvidia tax but still need the likes of TSMC, Broadcom, micron/sk hynix, Marvell, the truth semi-companies. CSPs will not have the patents, IPs and talent to replace any of them.

Also, not sure how well CSPs inference stack is compared with vllm + nvidia. A lot of open weight models uses MoE, making the inference stack more complex.

China does not want to crash US economy. US economy strong means more purchases for our goods. But US want to crash Chinese economy by: 2018 trade war, 2019 Huawei sanctions, sanctioning at least 1000 large influential Chinese companies, 2020 Tiktok theft, 2022 total ban on GPUs, semiconductor equipment going to China regardless of end use, banning all sorts of Chinese goods such as EVs, cars, batteries, drones, even DJI cameras, 2020 fake lie about Xinjiang to implement full ban of anything made by Xinjiang, 2025 50%+ tariff on Chinese goods. I don't think building LLM was thought has important until Anthropic went on full China hating, banning anyone on a Chinese IP using the service, even companies headquartered in China's oversea offices. Did you know that a regular person in China trying to use Claude to ask some innocent question gets banned? Talk about everyone created equal and democratic world.

You leave us with no choice but to build technologies ourselves. If its bad for Anthropic, well, you could owned the market in China, oh well, you reap what you sow.

You need 64 H200 super-node for inference for kimi k3. You will not do inference locally. What might pop the western hardware bubble is Chinese GPU, memory, networking companies. But even in China, these AI centric hardware is not cheap.

No, Nvidia is worried about Chinese hardware stack. China is under sanction, what are they training and inferencing these models on? Even if this model might still be Nvidia chips, what about the next model. Everyone knows model and hardware companies are working together. There are a number of very competitive companies in China in this space. After they got this area sorted out, China will do training, inference and tokens entirely on their stack and export their entire stack.

Winners are hardware companies, GPUs,XPUs, HBM, memory, connectivity. Even CSPs are just compute renters, they charge a margin to make sure their hardware purchases can be made back. But given there are more and more AI CSPs, traditional, and neocloud, pricing competition is inevitable, and given the huge expense of hardware, CSPs are squeeze by the hardware companies and users seeking to lower their own costs.

Chinese solar technology is multiple generations ahead of US companies. First solar panels have a conversion rate of 17% while longi is 24.8% and demonstrating 26%. First solar should be out of the market without US government subsidies and protectionism.

WS 15 is already shipping with latest J20, its trust to weight ratio is above 10.

CJ1000A is being certified by 2028. It is equivalent to LEAP engines on fuel burn. WS15 and WS19 are all shipping today and have thrust to weight ratio above 10. CJ2000 is a 35-ton trust, genx fuel burn class engine, on track for 2030.

if you look at the price per gb of memory, at today's prices its only equal to 2014. People don't complain about memory prices back then. The reality is modern world just need much more memory. Before 8GB memory was enough, today my 32GB memory is 95% used all the time due to chrome. And my 2TB SSD is filled with 4K HDR videos. This is not counting massive storage needs by datasets, AI models, KV caches, and bandwidth needed for LLM inferencing.

US restriction on China is not just GPU, but total blockade of anything semi related all the way from fab equipment to final chips. Same restriction will work on any other country. But it does not work on China. Not only there are astonishing number of crazy smart AI researchers in China, but also the entire semi supply chain, from fab machines to GPU/XPU chips and software ecosystems, is advancing extremely rapidly. China will be the only country where every step of the AI supply chain from materials, fab equipment, lithography, 7nm to 3nm logic fabs, HBM, packaging, photonics, GPU/CPU/XPU, software ecosystem, frontier AI labs, power production and power generation equipment are all within a single country.

Maybe software people can get around this issue by not making every app a electron bloat? This is now more than doable now you got AI right? And it will save your job

cxmt is also selling their memory similar to the big three. No one hardware company in their right mind will sell their products not as high as possible after you learn how much harder hardware engineer and fab people work for.

if demand is too high, the market will adjust by adding supply. If Apple can't take the price of memory, why don't they make memory themselves. Oh no, the technical and manufacturing know-how is a barrier to entry and there is no talent available to make it. That is memory or semi companies' competitive advantages and their pricing power. It also takes Phd degrees to work at hardware companies. You can't have people doing leetcode for 2 months and todo apps and getting into the field and make 300k a year and hardware engineers wih Phds making 100k before this boom. It is long overdue for a rebalance of pricing power between hardware and software companies

memory is a commodity is laughable. Then software engineering is even more a commodity, the amount of engineering going into making memory chips the vast majority of people don't understand. There are a lot of software engineers getting this field after leetcoding and copy from hellointerview. Claude can write you an app in 30 minutes. Try build a lpddr5 dram chip in 30 minutes. Manufacturing know how itself is a specialty and barrier to entry.

And hyper-scalers, Apple, AI labs all will die if memory makers can't sell to them?

Lol, software company execs didn't see this coming. Fire all your experienced devs to jump on Anthropic bandwagon. Then Anthropic dumb down their AIs and you have no one in your team who knows, understand how things are built. Your entire company goes down. Your entire company's operation depends on the whims of Anthropic. If Anthropic raises prices by 10% per year, you have to eat it. This is what you get when you don't respect human beings and human talent.

Chargers don't charge at 1.5MW 100% of time. You have grid storage batteries serving as buffer. It can be charged at steady rate by the grid all the time. People need time to drive in and out of the station. The math works out really well.

They simply use a few grid storage batteries. Chargers don't charge at 1.5MW 100% of time. You also have people driving in and out of the station. The math works out really well.

BYD is shipping both charging stations and cars with blade battery 2.0 first half of 2026. Both economic and premium models can charge from 0 - 97% under 12 minutes. They are also building these charging stations right now and few hundreds are already operational.

there needs to be battery chemistry improvements. Otherwise, with existing batteries, charging at these speeds will cause too much heat and shorten battery life span. BYD is offering 1.5MW charging with increased battery lifespan and without increasing the heat dissipation requirements. Another improvement compared to current crop of batteries is charge curve. Charging from 80 to 95%, BYD batteries can handle higher power than current batteries at MAX