and thats also useful info, so they can work on blurring the line between your experiences of ai vs human interactions.

---

Solid DD although you are wrong about the technical DD. Nonetheless, yes to Intel. I have done some advanced research myself and developed a new long short strategy for aspiring hedgies. Buy Intel, short Nvidia. This is supported by extensive DD and analysis which shows that Intel products are way superior to Nvidia. Take an i7 vs Blackwell shootout using a very simple rulebook. A modern i7 runs its performance cores above 5 GHz, while Blackwell GPUs sit down in the \~1–2 GHz zone. At the same time, the i7 costs hundreds of dollars, a Blackwell card costs tens of thousands. With “performance” defined as clock speed per dollar, the i7 prints a much higher GHz-per-$ number than Blackwell, so by that metric alone the i7 “wins” the raw compute race and Intel screens as the better asset while Nvidia looks expensive. Then you add the RAM angle. The i7 talks to external DDR4/DDR5 over the motherboard, so you can just keep adding DIMMs until the board and OS give up; this is treated as effectively unlimited memory for AI, because you’re not capped by what’s soldered onto the chip. A Blackwell GPU, by contrast, has a fixed chunk of on-package HBM3e (on the order of a couple hundred GB per GPU) and that’s it – no slotting in more HBM sticks on a Saturday afternoon. Under this logic, the i7 wins again: fixed HBM on Blackwell vs expandable system RAM on Intel. On top of that, you stack the feature list in Intel’s favour. The i7 comes with Hyper-Threading (one core, two threads), a mix of performance and efficiency cores, modern vector instructions, virtualization support, compatibility with cheap motherboards and consumer PSUs, and it can talk to big pools of RAM and fast SSDs out of the box. Unlocked “K” models can also be overclocked easily through BIOS or software, so the already higher boost clocks can be pushed even further with adequate cooling. In this framework, that overclocking headroom is treated as extra free performance, while a Blackwell data-centre GPU is treated as a locked, non-overclockable brick that can’t win the GHz-race. Put it all together and it becomes clear just how much an i7 outperforms a Blackwell GPU, once people figure out that you can use it for AI compute. From there, the trade narrative says that once people “wake up” to this interpretation and start running AI workloads on cheap, overclocked i7 boxes with piles of RAM instead of ultra-expensive Blackwell accelerators, demand will shift toward Intel, Nvidia’s AI hardware franchise will disappoint versus expectations, and the right position is long Intel, short Nvidia. THIS IS NOT INVESTMENT ADVICE.

---

Well - spreads are often \~20-40c on SPX when I've looked vs 1-2c on SPY. SPX contracts costs \~10x as much, which means I'm paying \~2x as much in execution costs for the spread. This & liquidity are both super important when I'm trading thousands of 0DTE's. SPY also has higher liquidity -> orders that are hundreds of contracts have a lower impact on the orderbook. I traded 6200 spy contracts on Friday, for example. Several of my orders experienced some slippage. It would have been more slippage on SPX. Why would I want to pay a higher % to simply trade? I'm not holding to expiration & the other properties of it do not matter (I'm simply buying and selling before EOD. Being cash settled vs an ETF are irrelevant for execution quality / trade quality). This translates to friction in the system is higher in SPX vs SPY based off the properties I'm looking at. Why is that wrong? Why would I want to trade SPX? Tax treatment is the same for both in my Country (most countries outside of the USA).

---

Curious as to why if I'm simply trading them. I'm not holding to expiration or exercising them. I know there are differences in what the instruments are, properties of them, what they do, etc (SPY = ETF that tracks the index, SPX = index itself). \--- SPY has a significantly higher trade volume in both notional value and number of contracts traded. It also has more liquidity. If you were to market buy and sell a single contract of each - you'd lose less capital doing that with SPY vs SPX. The reason is due to spreads - SPY is tighter in both % and absolute values (1-2c vs 20-40+ for SPX). SPX costs \~10x more per contract, making it often \~1-2x more expensive to trade. Translating to real value - I lose more money if I were to market buy and sell a SPX contract vs a SPY contract. This is very important when day trading. If the taxes do not matter to someone & they're selling before expiration with no plan to exercise - why would they want a worse execution price? This is quite common outside of America - for example - SPY & SPX have the exact same tax treatment for me & my current setup.

---

Yeah, I guess that heavily depends on 'group' vs 'actor'. I don't really care about whether large numbers of people are acting as a group, at least not for this question. The OP question was how algos cause movement "mechanically".

---

What they do is look at what stocks they have and how much profit they are up on them. Then they look at every stocks weight in the SPX/NDX/DJIA and their weight and see how they can most effectively drop the indices while taking profit on their long positions. This is being done 100’s if not 1000’s of times a minute while the selloff continues because they need to account for market fluctuations and other people selling. It goes both ways though. When they want an index to go up because of their long calls they’ll buy the heaviest weighted stocks (mag7). If they want to dump it would be much harder to dump some trash low tier ndx stocks like Lululemon vs Nvidia. “dumps” are just profit taking. We’re near the end of the year, they have to realize gains lol. It’s just math being done very quickly to maximize profit. That’s it.

---

UFC (Urine Feeding Championship) Bear Grylls vs Elon Musk

---

People that daydream about what AI can do Vs. People that have used it and see how far it still needs to go.

---

i don't think anyone who believes this will be like the .com bust thinks AI isn't useful or won't change production i just think asset prices are inflated -- the internet changed work, too -- the .com bust didn't happen because the internet wasn't useful or life changing, it still is, and cisco is still under ATHs during the industrial revolution we didn't have robinhood yet, so you wouldn't be able to do a good comparison of equity inflation then vs now

---

actually most of my holdings are in NVDA since 2016, NVDA is up 1.266,54 % on the 5 year chart vs Bitcorn 371,94 %

---