Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Strategies & Market Trends : ajtj's Post-Lobotomy Market Charts and Thoughts -- Ignore unavailable to you. Want to Upgrade?

To: ajtj99 who wrote (94499)	8/25/2025 6:14:11 PM
From: ajtj99	Read Replies (1) \| Respond to of 97584

I hope people read the S&P Global data I posted today. There is a lot of actionable data there for investors.

To: ajtj99 who wrote (94499)	8/25/2025 8:26:58 PM
From: Sun Tzu	3 Recommendations Recommended By ajtj99 Kahlua toccodolce Read Replies (1) \| Respond to of 97584

You are comparing apples and oranges in this post. DeepSeek has one of the most (possibly the most) efficient pipeline architectures. This was their true innovation, not their reasoning model (which by the way produces subpar results). However, any model, no matter how efficient, still benefits from running on better hardware. nVidia's hardware is unmatched. Trust me, I have tried to build similar systems with cheaper components by others and nVidia's still ends up being a better bang for the buck. So I concluded that I will just wait and buy their next release whenever it comes out (so far the estimate is for a 6 months delay from the initial estimate). Once that HW is out, I may very well run DeepSeek or Mistral or Gemma or GPT OSS or whatever. One has nothing to do with the other, except as a secondary effect that a very efficient model will require lighter servers to run. BTW, nVidia has set a target for photonic AI hardware in 2026. I am very skeptical of that, but even if they make it by 2028, it will be a huge success.