
TheValueist
@TheValueist
Followers
2K
Following
6K
Media
1K
Statuses
7K
L/S equity in tech and energy. ISO convexity. Path dependence matters. Sizing via Kelly Criterion. Results never lie. Not financial advice.
Above the Noise
Joined September 2024
Happy 4th of July!!! šŗšøšŗšøšŗšø. Watch this video in its entirety. Invest in all the companies who's products were mentioned in this video including merchant power and by July 4, 2026 you will be a very very very happy person.
Ever wonder what it takes to build an AI factory?. Hereās your all-access pass into how Dell brings AI to life - at scale, at speed, and around the world. š„ Watch the full @ServeTheHome tour:.
0
0
1
$META showed the same chart at Hot Chips today when presenting the configuration of their $NVDA GB200 racks.
Catching up on some old stuff. TBC this is a few wks old from $meta at OCP and we could argue w the Y-axis precision here, but in case you has any doubt whatsoever that Zuck is going for it š
2
0
2
Hot Chips 2025. This is the epicenter of the semiconductor engineering conversation globally. All about massive GPU clusters. The GAI revolution is only getting started. $NVDA $AMD $TSM $AVGO
$NVDA Hot Topics at Hot Chips: Inference, Networking, AI Innovation at Every Scale ā All Built on NVIDIA
0
0
0
$TSLA This is the worst it will ever be. Tesla FSD V14 to 'Feel Alive,' Second Biggest FSD Update Ever via @NotATeslaApp.
notateslaapp.com
A breakdown of FSD v14: how it aims to feel āalive,ā reduces driver monitoring, and pushes Teslaās AI forward.
0
0
0
True. $GOOGL $TSLA.
It seems like there is not enough of a policy response to the fact that, with 57M miles of data, Waymoās autonomous vehicles experience 85% less serious injuries & 79% less injuries overall than cars with human drivers. 2.4 million are injured & 40k killed in US accidents a year.
0
0
0
The Data Center Coalitionās filing to North Carolina regulators is a targeted attempt to shape how the state integrates very large new loads without imposing a bespoke ādata center tariff.ā The document asks the NCUC to keep service under existing tariffs with 3 pillars:.
Data Center Coalition (representing 35 leading data center owners, operators, and large-capacity lessees) urges NC regulators: Donāt create new ālarge loadā tariffs for data centers. Instead, use existing rate structures with cost-based pricing. Want flexible terms: 3-5yr load
0
0
3
RT @JeffWeniger: At $4.28 trillion, Nvidia's total market value is about a trillion dollars shy of the combined value of Japan's Nikkei 225ā¦.
0
147
0
$NVDA $GOOG $AVGO $AMD The exchange you shared is about when and why teams can avoid pipeline parallelism in largeāscale language model training if the hardware domain has extremely high insideātheābox bandwidth. The core point is that TPUs and large NVLink domains supply so much.
I got a bit lost in all the acronyms and jargon here, so I had Claude explain it without using any acronyms and now it all makes perfect sense (tldr; bandwidth ā¹ simplicity):. This is a fascinating technical discussion about training large language models at scale. The Core
0
1
2
$NVDA $TSM Hereās the translation of the article:. **NVIDIA CEO Jensen Huang made a quick visit to Taiwan yesterday by private jet. (United Daily News file photo)**. NVIDIA CEO Jensen Huang made a quick one-day visit to Taiwan yesterday (22nd), marking his third visit to Taiwan.
Big time: $NVDA CEO Jensen Huang: GB300 has now entered full-scale production, with output ramping up successfully. Full quote (translated from media): "Blackwell Ultra GB300 has now entered full-scale production, with output ramping up successfully. He emphasized that TSMC and
1
0
0
$NVDA Hot Topics at Hot Chips: Inference, Networking, AI Innovation at Every Scale ā All Built on NVIDIA
blogs.nvidia.com
AI reasoning, inference and networking will be top of mind for attendees of next weekās Hot Chips conference. A key forum for processor and system architects from industry and academia, Hot Chips ā...
0
0
1
$NVDA Inside NVIDIA Blackwell Ultra: The Chip Powering the AI Factory Era | NVIDIA Technical Blog
developer.nvidia.com
As the latest member of the NVIDIA Blackwell architecture family, the NVIDIA Blackwell Ultra GPU builds on core innovations to accelerate training and AI reasoning. It fuses silicon innovations withā¦
0
0
0
The core claim in the piece is that much of the AI āapp layer,ā particularly coding tools, is being misvalued because observers are conflating commodity throughput with durable software revenue. The author argues that token flow is not ARR and proposes.
growell.fyi
Throughput is not ARR. Value the app layer on GTV as scale and real margins, not on token throughput.
0
0
1
I 1000% agree with this. Gatekeeper lead generation business models will become extinct in short order. $NVDA.
0
0
3
Working on and in production are two very very different things. $NVDA.
OpenAI CPO, Kevin Weil:. We're working on our own chips and using AI to improve chip design and layout. "we'd be crazy not to". It's a well-defined, measurable problem (chip speed), so we can iterate with lots of GPUs and keep improving
0
0
0
$META.
1/ Today weāre proud to announce a partnership with @midjourney, to license their aesthetic technology for our future models and products, bringing beauty to billions.
0
0
0