Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
TSLA Cybertruck/ RIVN - IMO CT range is a big disappointment relative to original estimates 250-340 miles. I did put in an order soon after reveal and spreadsheet says late next year projected for my delivery. The original projections were 250-500 miles depending on configuration. My neighbor a few doors down put in an order minutes after the initial specs were provided. He wanted 500 miles so he can make it to his winter place without charging stops. I wanted at least 300 miles. This is not going to work. Rivian will be getting a lot of orders and will do well if they do better on manufacturing scaling (all sorts of different size bolts in early lots).
Kerry Emanuel of MIT is the foremost expert on hurricanes and he addressed the data in the link you provided. He gives an evaluation of that data in his UCLA presentation. It's important to understand how the data was obtained and the differences between "data collected" a century+ ago to decades+ ago to more recent methods, ie satellites and planes flying through eye walls and dropping sensors in and through the eyes.
It's obvious that the oil companies are just following the most profitable path for them. That is free market capitalism. However, they need to stop acting like the tobacco companies advertising that smoking is good for you. Their sponsorship of misinformation on "moral" grounds is laughable, unnecessary and
downright counterproductive. Lee Raymond has already done enough damage during his time at Exxon.
RE NASA data set - reports by ships vs islands vs satellites vs airplanes flying into the eyes.
The rapid rise over the last few years in homeowners insurance costs due to increasing risks of wild fires and increasing hurricane intensity for the states of CA, TX, LA, and FL is just the beginning of the costs of CO2 emissions. Perhaps the oil companies would like to address those acute costs, and costs of "managed retreat" for populated coastal regions around the world. My personal delta in insurance coverage over the last 3 years? Around 100% in annual premiums and that's with a policy deductible raised from $2K to $10K.
Kiwi - not sure what your understanding of PLTR's software role is at NHS. They have worked with NIH for a while now but how exactly their software is used only the imbedded folks know. NHS use of consultants like NECS, PwC etc using msft Excel/Azure really hasn't done good enough for PLTR to win. Now those consultancies will be Palantir shops for institutions that need something beyond fancy spreadsheets. I don't even know if this is just Foundry since AIP was non-existent when this bidding process started. IMO NHS contract is a nice feather in the cap but ultimately, foundry/aip will have to compete in the private sector where the large TAM is and where growth will have to come from.
Nice graphic of global electricity generation energy source. Other than Singapore, the countries with grey color trending up are places that you probably don't want to live in.
https://www.nytimes.com/interactive/2023/11/20/climate/global-power-electricity-fossil-fuels-coal.html
How things come full circle - Note from Mike Grossman -
Mike Grossman
2:15?AM (6 hours ago)
to Mike
Non-Newtonian calculus is used in the doctoral dissertation on artificial intelligence by Michael Valenzuela at the University of Arizona. The dissertation is called ”Machine learning, optimization, and anti-training with sacrificial data“. (In computer science, machine learning is a branch of artificial intelligence.) It includes sections called “Non-Newtonian Derivations”, Non-Newtonian Models”, and “Applications of Non-Newtonian Calculus”.
From the dissertation: “Traditionally the machine learning community has viewed the No Free Lunch (NFL) theorems for search and optimization as a limitation. I review, analyze, and unify the NFL theorem with the many frameworks to arrive at necessary conditions for improving black-box optimization, model selection, and machine learning in general. I review meta-learning literature to determine when and how meta-learning can bene?t machine learning. We generalize meta-learning, in context of the NFL theorems, to arrive at a novel technique called Anti-Training with Sacri?cial Data (ATSD). My technique applies at the meta level to arrive at domain speci?c algorithms and models. I also show how to generate sacri?cial data. An extensive case study is presented. … For algorithms designed to operate on non-combinatorial problems, derivatives or second order assumptions are often exploited. However, both the derivatives and de?nition of second order depends on which calculus is used. Commonly classical additive calculus is used resulting in the common quadratic model and the classical derivative. This need not be the case. Grossman and Katz [Non-Newtonian Calculus] mention several alternative calculi including: geometric, anageometric, bigeometric, quadratic, anaquadratic, biquadratic, harmonic, anaharmonic, and biharmonic. … Non-Newtonian calculus has been used to derive optimization algorithms that perform better than traditional Newton based methods for Expectation-Maximization algorithms. However, Non-Newtonian calculus goes beyond simply being useful for optimization, it is useful for the other half of learning: modeling. The second order approximation using geometric calculus may produce the Gaussian curve … . The nth order approximation using bigeometric calculus produces an nth polynomial on a log-log plot.… Non-Newtonian generalized Taylor expansions produce nth order models, which are rarely polynomials. … Non-Newtonian models sometimes make sense to use. Non-Newtonian models follow from non-Newtonian calculi. … Here are a few rules of thumb for non-Newtonian models. If a meta-model is primarily concerned with learning probabilities, non-parametric distributions, or anything else where the multiplication is the primary operation, then the geometric calculi may be of interest. If working in a domain where the squares are additive, as is common the case when estimating the variance of a sum of independent random variables, then the quadratic calculi may produce meaningful models.”
https://www.researchgate.net/publication/301692116_Machine_Learning_Optimization_and_Anti-Training_with_Sacrificial_Data
8:56?AM (0 minutes ago)
to Mike
Hah. I wonder if ChatGPT and BARD implementation of NNC will result in a better answer about "What's NNC?" Nah - it will just be faster. LOL.
I didn't know there was another option. Any idea what the other 5% is?
Yes. I guess XOM didn't want to increase single country/basin so they passed on HES. Good fit for CVX so both did well.
Interesting. Stanford peds insists that family members of newborn (esp preemies) get RSV vaccine when available.
XOM - pxd timing good as they can use a new person to head up shale operations.
https://www.thedailybeast.com/exxonmobil-head-of-shale-oil-and-gas-david-scott-arrested-on-sexual-assault-charge
Not sure what reality is but it certainly not representative of the Norcal area I spend time in currently. Tesla's everywhere and not just in fancy neighborhoods. It is true though that Tesla's do get written off as totaled if they get frame damage (as ICE cars do too). Cheaper just to sell for recycle/scrap and replace. PowerFlex was charging 0.40/kwh last week during day time heatwave vs. $6ish/gal of gasoline. I think Solar is going to win no matter what along with EVs. Public library chargers were 0.23/kwh at the same time (palo alto).
Logical move. IMO COP up on the docket given the minimal opposition to XOM << PXD gobble. It will have to be SHEL or CVX.
TSLA - Could this have been FSD or just a Texan being too Texan?
https://www.carscoops.com/2023/09/roadworker-filming-only-cements-this-stuck-tesla-drivers-problems/
KFC/CSCO - Call perhaps should be named after CSCO since Chambers should have known better than to outsource to China some of the underlying tech that enabled Huawei. We have our NSA, China is just copying that too.
https://arstechnica.com/security/2023/09/china-state-hackers-are-camping-out-in-cisco-routers-us-and-japan-warn/
MBGAF level III below 40 MPH ready for prime time. Unlike TSLA, Mercedes will take legal liability when their system is used. System developed jointly with Bosch.
https://www.theverge.com/2023/9/27/23892154/mercedes-benz-drive-pilot-autonomous-level-3-test
MSFT/PLTR - Fabric is MSFT answer to Palantir's Foundry and it will be formidable competition.
https://learn.microsoft.com/en-us/fabric/get-started/microsoft-fabric-overview
TM - All good but what about the batteries? We need TM around to compete. Also Tesla going with near full body casting for their low end model in effort to hit sub $30K market. To be built in Texas, then Mexico. It's a tough road ahead. Also battery packs probably going hybrid (lithium and sodium) IMO.
NVDA - Until proven otherwise, the rise of AI benefits hardware providers relative to software. BTW - I have found it useful to turn Bard on in my Google searches and get additional context on topics/things that search.
China EVs/economy - Does the absolute number of manufacturers matter? Does state of de-leveraging in real estate market matter? What do you think? Just asking.
WSJ equates China's 500 EV manufacturers to how many in the US? So China is down to maybe 100 brands after 400 drops out. Anything that Murdoch puts out are opinion pieces unrestrained by facts.
Waymo!!! - NYT reporters tested 3 different routes with 3 reporters in SF. Smooth and unaggressive is good. I'm ready to try it next time I'm in SF. Cruise is not ready for prime time.
https://www.nytimes.com/2023/08/21/technology/waymo-driverless-cars-san-francisco.html
LLM and prime numbers -
>>>>>>>>>>>>>>
Artificial intelligence, hold the intellect.
Is 1 a prime number? For the answer, you might want to ask Google (or a sixth grader) rather than ChatGPT. In March, two researchers from Stanford and one from Berkeley quizzed the AI model’s newest version, GPT-4, and found that it correctly identified prime numbers 97.6 percent of the time. A few months later, when they quizzed it again, it correctly identified prime numbers just 2.4 percent of the time. So what happened? Did ChatGPT get distracted by the cute AI model next door? Did humans literally dumb it down? The researchers say it’s complicated. “It’s very difficult to say, in general, whether GPT-4 or GPT-3.5 is getting better or worse over time,” said James Zou, an assistant professor of data science and a co-author of the preprint study.
It may appear that the AI is getting dumber, but the researchers point out that it also provided fewer potentially offensive responses and was less likely to offer ideas on how to break the law. Plus, they say, it wasn’t actually “smart” to begin with. Since ChatGPT is a large language model trained to generate human-sounding text, it might, in fact, never have assessed primeness at all but rather produced answers based on incidental trends in its training data. ChatGPT’s parent company, OpenAI, doesn’t discuss how it trains ChatGPT, so as it adds new training data—which could change the way the model responds—researchers and users are left to speculate. “While the majority of metrics have improved,” the company said, “there may be some tasks where the performance gets worse.” In the meantime, the Loop can help you out with this minor mathematical mystery: 1 may be lonely, but it’s not prime.
Polysilicon wafer prices are quite volatile and dropped a lot recently. The 1.5 cents/kwh include storage apparently. Without storage the bids are close to 1.0 cents/kwh in SA. The half cent delta is consistent with Arizona projects with and without battery backup,
Wafer and module spot/historical prices available via OPIS.
https://info.opisnet.com/hubfs/Product%20Campaigns/APAC%20Solar%20Weekly/OPIS-APAC-Solar-Weekly-Report.pdf
I use Saudi Arabia as the reference point for tracking cost trends since cost records there are easy to track since middle east deserts have always set the yearly low cost record. Arizona desert is about a year behind as of a few years ago. This is about utility scale solar, not residential.
TSLA - Regarding the boiler room operational call center that has been set up in Las Vegas to specifically handle range complaints. Supposedly no longer doing that so perhaps handling complaints about FSD instead.
https://www.reuters.com/investigates/special-report/tesla-batteries-range/
That's funny. I haven't checked solar costs/kwh in a while. The last time I checked it was starting to cross under 3 cents/kwh. It's now crossing 1.5 cents/kwh. IMO fossil fuels are on a collision course with economics. If economics dictates realities, then IMO "powerlineblog" should have something that replaces "powerline". Competing against a cost curve that continues to trend down for the foreseeable future is rough.
PLTR Q2 slide deck. Haven't listened to webcast yet but I understand there was a lot said about AIP adoption. If that's the case, AI is a low margin business for now and perhaps in the longer term. ML will just be part of the software fabric.
https://investors.palantir.com/files/Palantir%20Q2%202023%20Business%20Update.pdf
That's interesting. I have to look into who's doing the cruise control in my car. Clearly, it is much better than last year. What you described is like what I experienced last year - having to constantly override the control (by gas or brake pedal) and much less so this year.
Same sentiment as yours. It's all marketing hype just like "DOJO" is now. FSD is going to take AI at the edge (ie at the car), not some computer in an air conditioned warehouse. I have some Level II'ish cruise control in my current car for stop and go traffic and I'm sure TSLA is ahead. I also recently saw the latest reincarnation of Waymo Jag. One would think the Lidar unit will shrink with time, but no, it's probably the biggest Lidar rooftop unit I have ever spotted. Not happening anytime soon for non-geofenced approach IMO. Even for geofenced app like SF, population there is revolting and putting cones out to impede those vehicles.
TM/TSLA - California market share report. BEV in NorCal is around the same as China. SoCal lagging about 5% behind @ 20.9%. But I think the trend is clear, TSLA will overtake TM in CA market share soon. No rest for those that are slow and weary.
https://www.cncda.org/wp-content/uploads/Cal-Covering-2Q-23_FINAL.pdf
OpenAI/GOOG/MSFT - Bilingual Japanese journalists evaluation of translations from 4 LLM's. BTW - since reveal of the hack of msft's cloud, Windows startup now include Gdrive on taskbar. Still getting Windows patches for updates as of a couple of days ago.
https://www.japantimes.co.jp/life/2023/07/18/language/japanese-english-ai-translation
It's really not that hard. It's about gas properties. their refraction indices and simple thermodynamics.
https://www.academia.edu/33161453/The_Properties_of_Gases_and_Liquids_4th_Edition_R_C_Reid_J_M_Prausnitz_and_B_E_Poling_
Serious climatologists better start thinking about permafrost and gas hydrates. If recollection serves, there are gigatons of gas sequestered. Not all of it will be affected by warming climate and melting permafrost. But they are starting to see increasing frequency of gas blowouts in Siberia and Canada. This is not good.
https://www.wired.com/story/the-arctic-is-a-freezer-thats-losing-power
XOM-DEN good fit for their CCS campaign and EOR. Especially if they can get the process work better in shale.
Although the cost gradients are now strongly in favor of renewables especially in places lacking power infrastructure, it probably won't be fast enough to stop the positive feedback loop of warming causing release of sequestered CO2(ie forests, permafrost, ocean). Removing enough CO2 from the atmosphere will probably require engineering solutions. Some younger generation folks are starting to guess that we're screwed so just party on and hope for the best.
ZERO ~ percentage for hybrid market share in 10-15 yrs from now.
Not sure if "they" refers to the whole group, ie researchers AND investors. My guess is that the investors got it wrong since they dumped money into these academic type programs and didn't stop until it was overcrowded. Were both groups overdriven by the "Cancer Moonshot" initiative?
The ICE component In PHEV is the significant portion of the HP so it dictates the cycle time requirement. Same for HEV. If the hEV portion is the bottleneck in development than that would dictate cycle time. Since there is more complexity in HEV and PHEV compared to pure ICE or BEV, I won't be surprise if cycle time of combined power train is longer than either pure ICE or BEV. There is one model out there that has this flipped, the mx-30 that is suppose to use ICE purely as the range extender and having the EV portion as the main power/range engine. The EV only has a range of 100 mi and not competitive. It will be interesting to see what happens with the range extender engine but that portion is the bottleneck as the EV car has been out in the market a couple of years.
That's fairly consistent for ICE models in US, EU and Japan. China EV development cycle seems to be half of that but then batteries are off the shelf via CATL and BYD. There are now hundreds of EV models in China and a bunch of companies that China.ccp wants to stay afloat. TM cut BZ4x to 20K from 29K USD in china and still not selling. Audi doing almost similar % discount to move their lux EV,