Marcelo P. Lima

46.3K posts

Marcelo P. Lima banner
Marcelo P. Lima

Marcelo P. Lima

@MarceloLima

Trying to find moats riding S-curves. Entrepreneur, investor. It’s always Day 1. @IncreasingRtrns host. Sign up for updates: https://t.co/IXlHVhOANw

Miami, FL شامل ہوئے Mayıs 2009
3.8K فالونگ28.8K فالوورز
No Safe Words
No Safe Words@Cyber_Trailer·
Since my account is somewhat anonymous I’m going to disclose where some of the California high-speed rail money gets wasted. 99% of you don’t realize where giant chunks of the money is disappearing to. The California high-speed rail authority, literally owns thousands of parcels of land that are in various stages continued litigation, tenant improvements, eviction, and constant maintenance. For example, there are many homes and apartment complexes in the plant path that have been purchased years ahead of construction. Removing those tenants is a slow and expensive process. (let’s ignore the extra stress on housing that all of these destroyed properties are causing) In some cases, these are low rent apartments with a lengthy eviction process During that process, the state of California is the landlord and has to maintain the property codes the same as any other landlord. This means repairs, adding smoke detectors, fixing roofs, vegetation management, landscaping, paying off tenants to leave early, boarding up Windows, constant trash cleanups, towing vehicles etc. But the High Speed Rail Authority doesn’t just have to maintain these properties at normal cost. Every single bit of that work has to be done at California prevailing wage rates. The work can only be done through qualified contractors that have passed through a long series of idiotic mazes to qualify to perform the work. An average rate per hour (charge rate) for a worker to perform any service on these properties is approximately $200 an hour for labor only. The cost go up for specialized work, like electricians, plumbers, or machine operators. Properties that are literally worthless are being maintained at huge expense just so the next round of homeless transients can break into the property and cause more damage. For reasons I can’t explain, the process to finally demo and remove the structures takes years. I’m only mentioning the tip of the iceberg regarding my firsthand knowledge. Completely separate from those outlandish costs are the inflation caused by the construction. The prevailing word on the street is that nothing is getting done. The truth is that a lot is getting done and none of it efficiently. The amount of concrete being poured daily and monthly to build gigantic overpasses for both the rail and roadways is not understood. In these work areas, every concrete mixing company is fully scheduled out and cannot offer building materials for other basic services such as building a house often times for weeks when the average lead time for many of these services used to be one day. And that’s just the schedule, never mind the huge cost increases from straining the supply chain and Labor pool. The amount of concrete and steel that has gone into the structures so far is massive. Dozens and dozens of new water wells have been dug just for dust control. Thousands upon thousands of acres of highly productive tree fruits and nuts have been torn up and shredded. Utility scale solar fields have been uprooted and sometimes relocated at extravagant costs. Every type of business you can imagine has gone through either a closure, relocation, or a long-term tenant agreement with the rail authority. In some cases, it’s just a buyout where the business closes its doors forever. The owners get something all of the workers get nothing. Don’t get me started on how thick the layers of bureaucracy are for these minute tasks that occur on all of these properties. The inefficiency is far beyond your wildest dreams. In many cases, this is not related to fraud in any way it’s just absolute ignorance, red tape, and failed leadership. I can go much deeper into specific examples, but I think that gives some of you an idea of what’s actually happening in California. If a rail is ever usable, some portions of the structures will be decades old and already in disrepair.
English
212
736
2.9K
223.2K
Marcelo P. Lima
Marcelo P. Lima@MarceloLima·
AI can read every medical journal, watch every health podcast, and learn about every advance. Amazon is putting this in your pocket and it's going to be revolutionary.
English
0
1
1
870
Nick shirley
Nick shirley@nickshirleyy·
NYC is now spending $81,000 per homeless person The average median household income is $81,000 for a New Yorker You work hard and receive nothing in return while the homeless receive your entire paycheck for free Stop the waste and fraud.
Anti Fraud@AntiFraudClub_

x.com/i/article/2034…

English
741
4.3K
17.4K
297.7K
Elon Musk
Elon Musk@elonmusk·
AI5 will punch far above its weight, because the entire Tesla AI software stack is designed to make maximally effective use of every circuit. We co-signed our AI software and hardware. Bear in mind that AI5, while it can be used for training in data centers, is primarily optimized for AI edge compute in Optimus and Robotaxi. There is still significant room for improvement. In the same half reticle and same process node, we think a single AI6 chip has the potential to match a dual SoC AI5.
English
265
511
5.9K
351.9K
phil beisel
phil beisel@pbeisel·
Tesla’s forthcoming AI5 uses a half-reticle design, which is crucial for yield. A reticle defines the imaging area of a lithography machine, fitting two chips per shot effectively doubles yield. This means the Tesla chip design team had to carefully manage die features, for instance dropping the older ISP (and classic GPU) to make room for more AI cores. By contrast, NVIDIA’s Blackwell fills nearly a full reticle, making it a single-reticle design. If Tesla hits its compute and efficiency targets with AI5 in this half-reticle format, it’s almost like cutting fab requirements in half. And this has a big impact on Terafab, especially if it carries forward for AI6, AI7, etc.
phil beisel tweet media
phil beisel@pbeisel

Terafab may be the most essential vertical integration Tesla has ever undertaken— and it is truly non-optional. It will take years to build and will test even Elon’s speedrunning abilities to the limit, but that won’t stop him from trying. The breakthrough likely lies in overhauling the overall facility’s cleanroom model. By moving wafers in sealed pods with localized micro-environments, the fab no longer needs a monolithic ultra-clean space. Elon’s line about “eating cheeseburgers and smoking cigars” on the fab floor isn’t silly, it’s the practical reality of a radically simpler, cheaper, faster approach that could finally change the economics of chipmaking. This is all forced by the brutal “pinch” in chip supply. Tesla must produce on the order of 100–200 billion AI chips per year just to saturate its roadmap. That volume powers: FSD cars & Robotaxis (tens of millions of vehicles needing AI5 inference for near-perfect autonomy), Physical Optimus (scaling from thousands today to millions per year, each requiring AI5/AI6-level compute), Digital Optimus (the new xAI-Tesla software agents for digital/office automation, running massive inference clusters), Space-based data centers (AI7/Dojo3 orbital compute for GW-scale training and inference beyond Earth limits). AI5 delivers the ~10× leap for vehicles and early robots; AI6 shifts focus to Optimus + terrestrial DCs; AI7 goes orbital. No external foundry (TSMC, Samsung, etc.) can deliver that scale or timeline— hence the Terafab launch. Without it, the entire robotics + autonomy future hits a brick wall. Terafab isn’t optional; it’s the only way forward.

English
58
184
2.1K
339.6K
Retard Finder
Retard Finder@IfindRetards·
This is a level of retardation not seen since Biden.
English
415
1.9K
10.5K
206.2K
Marcelo P. Lima
Marcelo P. Lima@MarceloLima·
Commoditize your complements: Nvidia's Nemotron 3 Super 120b is free and can be run locally. Furthers the thesis that for many agentic workloads, free models will be more than enough.
English
1
0
0
842
Handre
Handre@Handre·
Handre tweet media
ZXX
283
2K
14K
549.6K
Marcelo P. Lima
Marcelo P. Lima@MarceloLima·
Tomasz Tunguz has an interesting post today showing that a free open source model running locally outperforms a paid one running on the cloud. This augurs well for the thesis that, for a lot of use cases, models are indeed commoditized and COGS for agentic SaaS will stay low.
Marcelo P. Lima tweet media
English
0
1
6
599
Marcelo P. Lima
Marcelo P. Lima@MarceloLima·
Pedro Domingos has been warning about Mustafa Suleyman since at least early 2024. If only @satyanadella had listened, it would have saved him a big headache.
Marcelo P. Lima tweet media
English
1
2
7
1.7K
Marcelo P. Lima
Marcelo P. Lima@MarceloLima·
Note: I didn't adjust for the $1 bn Figma termination fee, thus the step-up in EBIT from 2024 to 2025
English
0
0
1
618
Marcelo P. Lima
Marcelo P. Lima@MarceloLima·
$ADBE Adobe now trades at 11.4x EV/EBIT If this were the 80s the LBO firms would be all over this
Marcelo P. Lima tweet media
English
4
1
51
5.1K
NVIDIA AI Developer
NVIDIA AI Developer@NVIDIAAIDev·
🙌 Andrej Karpathy’s lab has received the first DGX Station GB300 -- a Dell Pro Max with GB300. 💚 We can't wait to see what you’ll create @karpathy! 🔗 #dgx-station" target="_blank" rel="nofollow noopener">blogs.nvidia.com/blog/gtc-2026-… @DellTech
NVIDIA AI Developer tweet media
English
118
252
4K
1.1M
Marcelo P. Lima
Marcelo P. Lima@MarceloLima·
The value investors who own TerraVest now have to think about what to do given it trades at higher multiples on NTM than businesses like Microsoft and Meta.
English
2
1
6
2.6K