[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]

@pbeisel "NVIDIA stock rose 1467% from a low of $XXXXX on October XX 2022. Thats roughly a 14.7x increase over about X years and X months (1010 days). AMAZING $NVDA What might happen with $TSLA @CernBasher has some ideas. He shows $TSLA reaching $2300 a share in 2030 after he discounts his model. Robotaxi would make up around XX% of Tesla's profits at that point. That's about a 7x rise in the next X years"
@pbeisel on X 2025-07-20 01:18:56 UTC 14.9K followers, 6991 engagements

"@badgernewman It always was going to be a long year before seeing profits from Robotaxi. Nothing has changed other than what I heard is they are going to scale quickly. Maybe quicker than most expect"
@pbeisel on X 2025-07-24 00:40:09 UTC 14.9K followers, XXX engagements

"@TOC4AI @herbertong @grok I was so disappointed in Apple Intelligence. I've never seen Apple introduce a product that just doesn't work (except the Newton which sadly I worked on)"
@pbeisel on X 2025-07-20 16:58:02 UTC 14.9K followers, XX engagements

"Tesla Q2 2025 call and financials: X. Overall the financials weren't bad if Wall Street focuses on the thing they claim they care about they will be pleased. X. Robotaxi news bullish on track for rapid expansion this I care a lot about its the autonomy transition point for the company and they are clearly focused on it. X. AI X perhaps pushed out but honestly it doesn't matter at all they are doing the necessary with HW4. X. Optimus: wet blanket thrown on near term news on Optimus Elon bullish about version X less bullish on building the damn thing. I am not worried plenty of time to get"
@pbeisel on X 2025-07-23 23:18:46 UTC 14.9K followers, 56.6K engagements

"$TSLA went on sale today so I bought some"
@pbeisel on X 2025-07-25 04:05:46 UTC 14.9K followers, 11.5K engagements

"On Robotaxi: its clear to me from yesterday's call that the team is highly confident in the technical solution. In fact I'd say we are largely moving past the technical risk phase of Robotaxi and moving into the operational risk phase which includes regulatory approval. There is a lot to do operationally including expanding their skeleton Robotaxi operations team. Now many will say "but there are still safety observers". Yes but read the room they are seeing past this. They indicated they have 7000 largely intervention free miles under their belt. There is confidence here if you are listening"
@pbeisel on X 2025-07-24 15:53:32 UTC 14.9K followers, 16.2K engagements

"@Mdfgator91 In the linked post I said: "Deciding that version XX will be 10x bigger than version XX implies the team is confident they can run it on HW4 Tesla's. In other words they know the performance can be achieved for inference." Elon replied to that post with a 🎯 and reposted it"
@pbeisel on X 2025-07-26 21:27:26 UTC 14.9K followers, XXX engagements

"@fritzstauff Elon said 10x on earnings call so we will go with that"
@pbeisel on X 2025-07-25 02:44:39 UTC 14.9K followers, 5069 engagements

""Robotaxi will likely reach half the population of the US by EOY" -Elon Musk (7/23 earning's call) Assuming regulatory approval. He is obviously talking about hitting cities where the population is"
@pbeisel on X 2025-07-23 21:38:02 UTC 14.9K followers, 11K engagements

"10x Model and HW4 / AI5 We heard Musk speak about AI5 (aka HW5) on the Q2 '25 earnings call. He said it would go into production end of next year". That timeline tells us something important: Robotaxi does not require AI5. In other words Cybercab itself does not depend on it. The current Robotaxi fleet of Model Y's in Austin is running HW4 with version XX builds. And the next-gen model with 10x the number of parameters (aka v14) is expected to run on HW4 as well (see post below). That means Tesla can scale Robotaxi deployment without waiting for AI5. I do believe AI5 will be required for"
@pbeisel on X 2025-07-26 21:20:24 UTC 14.9K followers, 16K engagements

"@PaulWells69 Calibration. Your brain can estimate a distance to an object reasonably well but a tape measure will give you a precise distance then you will be well calibrated to that space (other measurements). Read the patent for more info"
@pbeisel on X 2025-07-22 17:55:26 UTC 14.9K followers, 1878 engagements

"@mehauff7 Thanks. No Waymo's approach is non scalable and will prove a big mistake. Tesla can validate where they wish to and then scale to the moon"
@pbeisel on X 2025-07-22 14:32:26 UTC 14.9K followers, 1478 engagements

"Different problems require different model architectures and sizes. xAI is building a large language model (LLM)or more accurately a next-generation general intelligence model. Understanding and generating language at that level requires learning associations across an enormous high-dimensional spaceessentially encoding relationships across entire books contexts and reasoning patterns. FSD on the other hand is solving a more constrained domain: driving. The model must process a finite set of input modalitiescamera frames vehicle dynamics map priorsand make real-time decisions. While the task"
@pbeisel on X 2025-07-25 02:59:47 UTC 14.9K followers, XX engagements

"My parents had their first date at a Drive-In (The Circus at the Jersey Shore). Its American history and future again"
@pbeisel on X 2025-07-23 17:28:25 UTC 14.9K followers, 2483 engagements

"10x Elon said the next version of FSD presumed to be version XX will have 10x the number of parameters as version XX (including the version XX builds used in Robotaxi). What does this mean Architecturally a neural network is made up of layers and those layers contain nodes (also called neurons). Each node in one layer is (typically) connected to every node in the next layer. These connections have numerical values called weights which determine how influential each connection is. Each node also has a bias which adjusts the nodes output independently of its inputs. The parameters in a neural"
@pbeisel on X 2025-07-24 17:23:04 UTC 14.9K followers, 2.4M engagements

"@Mdfgator91 Yea I'm not talking about market perception I am talking about my read of the room. What they are thinking. The market will lag this significantly"
@pbeisel on X 2025-07-24 17:00:58 UTC 14.9K followers, XXX engagements

"@CernBasher When does the Uber safety driver come out"
@pbeisel on X 2025-07-26 14:27:59 UTC 14.9K followers, 3079 engagements

"@denatalguy1975 Why improve anything They built off of XX for RT launch and now they are training XX just makes sense"
@pbeisel on X 2025-07-24 17:02:08 UTC 14.9K followers, XXX engagements

"@garrytan I put into typos but the LLM takes them out. 😆"
@pbeisel on X 2025-07-26 00:37:28 UTC 14.9K followers, XXX engagements

"@Austen "Prompt Engineering" is so stupid"
@pbeisel on X 2025-07-23 17:42:54 UTC 14.9K followers, XX engagements

"@ScottPresler from California with love. only in California. (yes this is an particularly expensive gas station) (no I don't give X shits I own EV's)"
@pbeisel on X 2025-07-26 18:58:09 UTC 14.9K followers, 4254 engagements

"@ConsensusGurus So let him sign the books. He is a brilliant guy. Should not speak to Wall Street. Are you reading what I wrote"
@pbeisel on X 2025-07-24 00:04:18 UTC 14.7K followers, XX engagements

"FSD Miles This graph is about to go straight up"
@pbeisel on X 2025-07-23 20:27:21 UTC 14.9K followers, 13.8K engagements

"@petergranthay @EVRobes @CernBasher Yes but if he is going to open his mouth. This is the dude that Wall Street is listening very carefully too"
@pbeisel on X 2025-07-24 15:57:54 UTC 14.9K followers, XX engagements

"About Me Im writing this after receiving a particularly nasty DM from an X user who made some mean ridiculous and incorrect assumptions about me especially regarding my positive posts on Tesla. I dont know what motivated him but I do know what motivates me. X. Im a Tesla investor. I own a significant amount of $TSLA and try to buy more anytime it goes on sale (like today). Im not chasing short-term gains I tune out headline noise and I honestly dont plan to sell any before 2035. I dont tradeIm not good at it. My edge is knowledge a long-term mindset and a high tolerance for volatility/risk."
@pbeisel on X 2025-07-24 19:09:37 UTC 14.9K followers, 43.3K engagements

"More better faster. Same reason ultimately you buy a newer iPhone. AI5 will reduce the photon to control latency further and support much much larger models. This just makes FSD safer and most fluid. Also AI5 will allow for other inference to be done on the car at the same time as FSD like a local Grok. But also AI5 is the brain for Optimus pretty much a requirement here"
@pbeisel on X 2025-07-27 00:15:22 UTC 14.9K followers, XXX engagements

"@WuWeiSci This 'validation' step will be removed as the measured deltas they see diminish. It will go away I just don't know when"
@pbeisel on X 2025-07-22 18:01:38 UTC 14.9K followers, XXX engagements

"That would make more sense but 12x12 seems big for a wafer. I'd have to research more on the size of the HW4 chip to see if this is reasonable. That said it matters little the DOJO configuration will use many of them for both inference and training use cases"
@pbeisel on X 2025-07-26 22:04:24 UTC 14.9K followers, XXX engagements

"@TeslaPrice XXX Sorry I don't understand your question"
@pbeisel on X 2025-07-26 21:36:19 UTC 14.9K followers, XXX engagements

"@HarrisonGrego20 @ScottPresler Yea its not Chevron's fault"
@pbeisel on X 2025-07-27 00:44:17 UTC 14.9K followers, XX engagements

"@theBocker70 Rivian and Tesla are very similar very 'first principle' companies. Both highly vertically integrated. Both run by their smart founders. Tech/R&D is done in silicon valley"
@pbeisel on X 2025-07-25 00:08:51 UTC 14.9K followers, XXX engagements

"@youyouAllen No I don't think so I believe the vehicle specific camera positions are input to the context window and relative offsets are computed on the way into to inference"
@pbeisel on X 2025-07-23 17:03:26 UTC 14.9K followers, XXX engagements