Search results

IMHO, China is happy to subsidize a company running in an oversupplied shrinking the margins for US manufacturers as this is the exact play they have used over and over.

Once my rental is paid off, I will be seeing where to spend some gambling money. I don't know if I have enough to move the needle for something this capital intensive, but worth asking.

Agreed. This is the same thing we have seen over and over and over....why is this time different? The dot com wasn't wrong. They just went too fast in the short term.

I am not willing to short them, but I 100% agree they're over priced. In fact, most of the market is overpriced.

My portfolio is doing amazing based on my thesis for AI. Time will tell if this holds.

I think the real value chain is LEO edge compute for FSD. An all electric fleet of tractor trailers with RT inner loop on board (controls are sub 5ms). For logistics planning, you have space edge compute that can see the full picture of traffic, routes, recharge stations, across both trucking...

The moon is ~ 239k miles. c is a vacuum is 186,282.4 mi/s. How much latency is acceptable here?

https://www.brightfarms.com/ I suppose it would not be technically impossible to produce some food in space, but food is extremely low margin so making that profitable would impossible. But this is obviously hyperbole used as a rhetorical device.

https://www.futuresource-consulting.com/the-source/industry-pulse/worldwide-dram-channel-market-faces-severe-supply-constraints-as-shipments-fall-to-historic-lows-in-q1-2026/ https://dramwatch.com/ The chart you uploaded shows smooth, monotonic bit growth from 2022 → 2027 for both supply and...

No, but at the same time the big 3 haven't increased supply, which supply demand dictates despite price increases. https://www.trendforce.com/presscenter/news/20260601-13070.html...

Global GDP gains about 3% average or about ~3.5T a year. Eating GDP isn't sustainable economically. This is just a supply demand law. So you have to GROW it. And of the 3.5T, I don't see AI taking that much of the growth.

I can't tell from your post or not, but FP64 is kind of the opposite direction from AI. CFD, FEA, Physics simulations are what you need FP64 for in order to maintain stable matrices in solvers.

I don't know about you, but the a100 is nearing my price range and when the h200 drops to 4k-ish effing yes I will buy one!!! That is a LOT OF FP64 compute!!!!

Well even if he is right (and I don't think he is), he is missing an important part of this. Namely that newer chips will give better performance per watt, and all the older hardware will be getting replaced at approximately the same time meaning cheap used parts. Used hardware is the space I...

Is that really true though? If all you're looking at is the average frame rate, sure. But there's a lot more nuance you can parse across the data. Lows, distribution of lows. Shader compilation time, sim games, etc...And, I would add running local models into the mix as well. I am not debating...

I am jealous for sure, however my 5950x over clocks to5.4gHz all cores and even in heavy axv2 loads it holds an all core clock of 4.8gHz and I don't like the cost timing is enough yet to make me feel the itch. After Zen 6 drops and after threadripper zen 6 is released, I am considering going...

The question of what to measure is indeed interesting. IMHO, there are reasonable debates on what should be measured and how. What I think is important is we are explicit about what is being measured and the limitations such measurements have in terms of performance inferences, which I think is...

Do we believe Nova Lake S makes zen 6 non competitive or what am I missing? Intel adding memory bandwidth, cores and cache vs AMD adding memory bandwidth (dual IMC), cache and cores and adding sea-of-wires. No idea how things will shake out, but I don't see this as a Core 2 moment for Intel.

cyankiwi/Qwen-AgentWorld-35B-A3B-AWQ-INT4 <--- okay...this is a whole new level of local ai now...just getting into this

Paul Kurgman isn't all economist...I can't stand the guy

Even when data center demand drops, you have companies waiting for price drops to put in orders. And then richer consumers. And then everyone else. Going to be a hot minute before we should expect prices to come down.

There's truth to this, but that doesn't change anything about how I am investing. If anything, that just means index funds will be barely over flat YOY for the unwind. Which is much worse than bubble with forced selling. In either case, until indexes match historic P/E, they don't make sense to...

Oh...well....shoot. I will just go mention to my farmer buddies that they can raise their prices and the people will just have to accept it. Foolish people assuming the market dictates a price based on supply and demand.....

Yep....I have diversified out of the big index funds in my 401k and personal accounts. I have 70% in dividend funds that mostly match SCHD...ish. The rest is sitting in treasury funds (effectively cash with ~3.5% returns...barely over 0 after inflation). This the dollar cost averaging mechanism...

Building a pension from working for years is the classic method, but current methods are more 401k contributions. Seems pretty weird for anyone to claim either of these are not earned benefits rather than UBI unless you're saying UBI is like contributions to a fund with a manager that then pays...

https://opnsense.org/

What is the 'too long, didn't watch' summary? Some videos have an AI summary, but I couldn't find it and I don't want to watch it.

Intent has nothing to do with anything. I intend to rule the world....so what? I make a tool to help me rule the world...so what? You need someone to operate the machine at the factory and the person and why have either? Per person production goes up. And IMHO, that is the real intent. And I...

People keep talking about AI like it’s some unprecedented force that will erase all human work. That’s the same mistake the Luddites made. We’ve never been able to predict the long‑term impact of a technology we’re seeing for the first time, it’s like asking someone to imagine a color they’ve...

I had a nice chunk in those and sold and made a nice chunk. I prefer asymmetrical upside or stability. So I am not going back in. I will stick with high value and dividend investing for the foreseeable future.

Repo markets are stressed right now and while the fed is holding rates steady, they are expanding their balance sheet. This is going to create more volatility. p/e ratios make no sense to me. What's a good sensible investment at this point?

I am up 26% this year in indexes and sold it all. Moving it to high dividend funds, not because I am a dividend investor, but because the companies in those funds are solidly profitable and for my own sanity I need to see that for it to make sense to me.

https://youtube.com/shorts/oh6EFElkUp4?is=T2qUJjR_nuP8wkL0 Feels relevant.

https://youtu.be/w8jiiVEW8r4?is=v26uifeADlY-ISub Interesting take here.

p/e ~ -233 at the moment

https://www.youtube.com/live/fwXwuMRfyDQ?is=_huGqUKNG0Sm6VfS Holy moly that valuation movement is insane. I couldn't handle the stress.

I am moving everything into high dividend etfs/funds at this point because they're priced correctly imo. If these future projections are right, those companies will be the beneficiaries and they're not priced for those forward gains making them a discount right now. If these future projections...

https://youtu.be/QYSmQx55e44?is=t76ReGdKhbjhrfpE Interesting. I can get a steady 65-70 tokens/sec on my 2x 3090s. That feels both good and bad. I have a much larger context window, which helps.

continue.dev extension works ish. But you won't get the copilot experience. Alternatively, you can use an mcp server and just have agents that do the lifting for you.

There's usually something that logs the token rate, cache hits etc...but not sure for ollama? There llm-benchmark on GitHub you could try. https://github.com/thomasdavis/llm-benchmark https://pypi.org/project/llm-benchmark/ there's a wheel for you too