nsx5200

Note that a lot of times in these type of problems, sometimes it's good enough to reduce the search space down to one where it's doable to brute-force/solve with other methods. Described more generically, these types of problems can be solved with a series of tools/method to arrive at the solution. Advancements in Mathematics are done this way, and this BTC 'problem' is no different than other mathematic problems. This potential 'flaw' is not present in physical gold, so to say that BTC is digital, better version of gold is not an accurate representation. Each has its own set of pros and cons.

It looks like my limits on SKYH triggered this morning and now I'm a potential bag holder of SKYH now. May the rich stay rich while using their private planes so us peasants can at least get some breadcrumbs. "Let them eat cake"

Short-term, pay the fee. Long-term, build your own pipeline and security to protect it. If Iran is smart, they would price their fee to match what it cost to maintain the pipeline, and periodically attack the pipeline to make the maintenance cost high. Pipeline owners would periodically allow these attacks to succeed so they can charge a premium for running material through their pipeline. Win for the pipeline/strait cartel without explicit collusion, which would invite the west to intervene. On another note. I'm surprised that the the solar industry is still down so much from the Trumpian policies. I would think at some point in time, reality kicks in and alternative energy generation like solar/wind must enter the picture to help fill in gaps.

Paypal thought they had enough customer facing leverage in 2016 to force Visa's hand but ended up settling. Companies like Visa/MC always had "threats" like this come up with similar justifications, but there's just too much legacy network effect moat and the value-to-cost provided by economy of scale to overcome. I have not seen any credible new justification on why Visa/MC would be displaced this time, and would appreciate any counter-argument that addresses the moats.

A bit late, but... Thank you for sharing your thought. It's a good problem to have when you need to decide between two good choices. I much prefer that than having to decide between two bad choices.

Apologies. @backtothebeach, do you mind explaining the logic behind this pair trade? I can't seem to figure it out. The best I can come up with is that you're swapping one value for a better value due to random walk of two stocks/variables. No real correlation involved between the two stocks in the classic pair-trade definition. Let me know if there's another angle that I'm not seeing. TIA.

In the US, Costco takes only Visa (Canadian Costco only takes Mastercard. Grr...). Costco is my primary basic necessity vendor so no go trying to move away from either one.

So it looks like the hackers are already on it. Data leak through AI LLMs even with all the guard rails in place. Like all the previous new 'features' before AI, everybody rush to adopt it, and then think about the security implications later (ex. weak WEP encryption in previous the first generations of WIFI). Not saying it can't be solved, but it does seem like this require a new set of techniques and time to shake out these 2nd order issues. In other news, I heard in the conspiracy grape vines that China is seeding content on the open web with AI slop that's favorable for the CCP in order to get future LLMs to train on this and become more CCP friendly. The long-term implication of this is that verified contents generated, or at least hand-checked by human will be valued more than output from AI systems, and rewarded accordingly in the future. If anything, preventing future LLM from digesting output from previous AI LLMs is a positive towards better performing LLMs. I'm starting to see a bit more articles from the main stream media start questioning these second order effects, so there seems to be some visible light at the end of this SaaSpocalypse.

See if they're putting money where their mouth is: https://www.anthropic.com/careers/jobs https://seo.ai/blog/how-many-people-work-at-anthropic IMHO, there's a disconnect between their marketing and their actual hiring practices based on this quick search.

Ran across this in my feed. This is a PhD/researcher's POV on the practical and theoretical limitations of current AI(LLM) systems. It talks about different cases where the models can break and degrade, especially if trained from output of LLM output. This has deep implication since a lot of these AI system's output is put on the internet, and the same content on the internet is used to train future AI systems. In addition, it goes into difference between syntax and semantics. The researcher claims, with evidence, that current AI LLM operates on syntax(words/symbols) without true understanding(semantics) and so there will be fundamental flaws in how these LLM tries to solve certain type of problems. For me, this was a bit of thought-provoking, as IMHO, many people, like these LLMs, walk through life operating mostly on syntax as well. Many of these syntax-based jobs will be replaced by these LLMs. Probably a bit more academic than most of those content produced by people with financial background, but IMHO, has more solid evidence, albeit a bit more theoretical, than the "I vibe coded, and it produced something somewhat working" evidence.

This Claude panic sure rhymes with the DeepSeek panic about a year ago. Mr. Market sure is bipolar. I also see a rise of 'hackers' that target these AI packaged answers and exploits the market when the AI packaged answers are incorrect. New areas for humans to explore, or could that be automated by AI?

This is almost a decade old, but I think it's still relevant: https://www.quora.com/What-would-happen-if-the-electricity-cost-to-mine-bitcoin-exceeded-the-value-of-the-bitcoin-reward "Bitcoin’s difficulty retargeting reduces mining difficulty after sustained hashrate decline (protocol adjusts every 2,016 blocks, ~every 2 weeks). Lower difficulty reduces electricity cost per expected reward for remaining miners." "While hashrate falls before difficulty adjusts, average block time lengthens (fewer blocks per hour). Transactions confirm more slowly, increasing mempool backlog and fee pressure." "Reduced hashrate lowers cost of attacking the chain (easier to mount 51% attacks). However, if difficulty subsequently falls and profitability returns, security restores." It seems like crypto mining compete with AI in electricity and silicon(compute) resources. I can see that if this AI binge continue to grow, it may be the trigger that cause the collapse of bitcoin(lost in faith), especially if Bitcoin's halving(scheduled in 2028) occurs during prolong electricity and silicon shortages. The bitcoin price($70k) is ~15% away from breakeven mining cost ($60k) before we see a steep drop off in miners. It will be interesting to see how the Bitcoin faithful's behavior will change at that threshold.

If you really want to get get a feel for AI capability on coding, you should be watching nerdy videos from coders about AI coding like this. A TLDR summary is that AI coding is still a skill multiplier, and does not allow randos to create production-level products, which AI maximus believe. I find that there's a huge gap between non-engineers and experienced engineers in their perception to create production quality products. IMHO, new products are announced all the time in the media, so typical people think that with a few weeks/months, you can just have a product from start to finish. Most don't hear about the failures unless somehow they go viral or do something extremely stupid. While in between jobs, I used to work part-time for a company that helps people implement their random ideas for some compensation. The amount of input resources that these people have in mind in order to realize the product is so fantastically low that it's not even funny. Needless to say, without proper resource, most ideas failed, even if the idea is actually viable. The company was also shady in that they would string the customer along just to suck their money without really telling them how unrealistic their expectation is. Companies see the short-term productivity gain, and they're letting the white-collar jobs go. My humble long-term prediction is that, at the corporate level, this is going to turn out to be like automation, where all the competent companies will adopt AI. So after it shakes out a bit, all the companies, relative to each other, are still going to be roughly the same, except more productive(especially if they can get rid of most management and overhead). All the barriers to entry will be roughly the same, with more edge for the bigger corporations due to economy of scale in acquiring and deploying resources. This will follow the winner-take-all trend that we've been seeing for decades now. Whether the white-color jobs will increase(Jevon's paradox) or not (job obsolescence) looks to be context dependent, but most are still TBD.

IMHO, there is also a time element to that as well. Like on Tariff Day, VIX spiked, and there's a near 100% chance that it will fall back down, the uncertainty is in the time. Using some system helps to identify the situation and the tool needed. Some situations requires more hands-on trading or constant monitoring, whereas others can more or less be left alone for decades. I actually started trying to keep track of the thesis, estimated return ballpark range and estimated timeframe in a spreadsheet (and not a DB w/ custom queries for those that read the Microsoft related thread), so I can go back and review how wrong I was. After a while, I think it can be turned into a catalog of scenarios and proper tools(mental models?) for tackling them. Isn't that how the pros in other fields really do it? Measure the error, adjust, repeat and hope to be >=1% better every time? In terms of predictability, I almost think you get a feel once you spend enough time in the topic. I'm wary of slapping a single hard percentage on stuff because it feels like false precision, but I think it helps to put put down the general ball park range, like 10-25% vs 50-70%, or if really new topics a wide range like, 10-90%. Same thing can be said for guesstimating the time frame as well.

https://www.wsj.com/lifestyle/workplace/ceos-say-ai-is-making-work-more-efficient-employees-tell-a-different-story-6613ce9d "The gulf between senior executives’ and workers’ actual experience with generative AI is vast, according to a new survey from the AI consulting firm Section of 5,000 white-collar workers." I find it amusing that the more distance from using AI on work, the rosier the outlook. There are tons of non-technical finance people on 'social media' that is giving opinions on how AI is going to disrupt everything, and betting accordingly...

Instead of handing out snacks, these airlines should just hand out Wegovy pills. Think about the savings in prepared meals: "Typical Airline Costs (Per Meal) Economy Class: $4 - $15 (or even less on budget airlines) Business Class: $25 - $50 First Class: $100 - $300 (can include multi-course meals and luxury items)" and the savings in the logistics of cleaning after meals .

A bit off-topic, and possibly could be moved into the China Macro thread. There seems to be a disconnect between that narrative and this: https://www.reuters.com/sustainability/climate-energy/villagers-shiver-chinas-north-government-gas-subsidies-shrink-2026-01-15/ "Villagers shiver in China's north as government gas subsidies shrink" Didn't do a DD on this, but several reasons that I can think of are: there is significant frictional cost in using electricity as a substitute for gas as a source of heat. poor grid or electricity distribution infrastructure? CCP doesn't feel like 'wasting' electricity by heating people's homes, but rather it goes toward their pet projects (AI, robotics, industry, etc...) If China banned bitcoin mining (without banning bitcoin) due to foreseeing the energy issue, then it would seem like there is some concern about their energy supply, and that electricity overabundance narrative might not be completely accurate.

The latest narrative is that "non-coders" will be able to crank out software in one hour when it used to take a team of software engineers one years to do. Even Microsoft is getting impacted, which has both the AI and software component. So it looks like anti-software is currently winning out. IMHO, this current hype cycle seems kind of like automotive FSD back then(has it been 10 years already?). Lots of high expectation, but will take a while to actually deliver. Even now, looking at some of the videos from actual coders seems to indicate that to use "vibe coding", you need a decent understanding of the limits of these AI agents. You have to be trained in scoping and prompting to get it do anything practical. Even then, nobody's really tackle the issue of debugging, which engineers know is the real skill differentiator. AI coding is definitely a skill multiplier, but definitely not a skill replacement (yet?). Better code tend to be more compact, so the amount of good to bad code is highly skewed toward bad code. If all those AI training is done on code that's on public repository (aka bad code), then I would claim that the future of human-centric software development remains bright. Most experienced software engineers know that over time, technical debt becomes the biggest cost/drag to any project, and very few managers, especially non-technical ones, can grasp the implications of that correctly.

Bought more CSU. It doesn't seem like the whole "AI is killing software development" narrative has softened yet, so the stocks of these potentially-impacted software companies have not bottomed out yet, IMHO.

I wish all political news were this entertaining: 1:10: "Maduro greeted spectators in the courtroom, saying, 'buenos dias'. Smart legal strategy by Maduro. If he speaks Spanish, ICE might bust in and deport him back to Venezuela just, out of habit."

Thanks for that. If you don't another pair of eyes looking at that, do you have the source to that info so I can go dig at it a bit? I have questions on what qualifies as "software" vs others (ai, service, etc), and how they quantify "generational lows". TIA.

Currently, the t-bill market is around $6T, so $180B is not significant to cause systemic failure yet. There's a built-in incentive for people to use dollar instead of stablecoin (opportunity cost in potential dividend/interest), and the people seem to be using stablecoin to ease crypto transactions, similar to a Treasury Zero CI. I suspect in a panic, the Fed can step in to purchase just the Treasuries to prevent spillover into the traditional finance world while letting the cryptos runs its course.

Here's the original 2025 Nov. S&P500 article on their assessment of Tether: https://www.spglobal.com/ratings/en/regulatory/delegate/getPDF?articleId=3486415&type=COMMENTS&defaultFormat=PDF S&P Global Ratings has reassessed the ability of Tether (USDT) to maintain its peg to the U.S. dollar to 5 (weak) from 4 (constrained)." Apparently, it's lost the peg for a short period in in 2022: https://www.ft.com/content/5887ef43-d43a-4608-a1ac-aacc99f076b9 "Tether tumbled as low as 95.11 cents in European trading, far below the $1 peg that it seeks to maintain as it faced an intense bout of selling pressure." "But in an interview with the Financial Times, he[Paolo Ardoino, CTO back then, CEO now] declined to give details about its $40bn hoard of US government bonds because he did not “want to give our secret sauce”." It doesn't smell like a fully trustworthy company.

Google AI Overview: "Key Mechanisms for Peg Stability: 100% Asset Backing: For every USDC token in circulation, Circle holds a corresponding value in U.S. dollars or highly liquid, short-term U.S. Treasury securities. Redeemability: Circle guarantees that users can always redeem USDC for U.S. dollars on a 1:1 basis directly with them, creating a strong incentive for stability. Arbitrage: If USDC trades below $1 on exchanges, arbitrageurs buy it cheap and redeem it with Circle for $1, driving the price back up. Mint/Burn Mechanism: To mint new USDC, Circle receives funds; to remove USDC, tokens are "burned" (destroyed), maintaining the balance with reserves." The biggest risk that I can see with these "stablecoin" is that if something, internal or external to these issuers, that cause the public to lose trust in the convertibility of the coin, it'll create a run to redeem it for USD. Depending on how liquid their backing is, not every stablecoin can be redeemable for the corresponding USD. So despite being a "stablecoin", it can, in times of panic, not be so stable. These coins are not riskless equivalent of the USD. Not saying there aren't risks to USD, but there are additional risks and features to these stablecoins.

For those that like Freakonomics, there's an old 2018 podcast on this as well: https://freakonomics.com/podcast/nurses-to-the-rescue/ "The A.M.A., as you might imagine, is not the biggest fan of retail health clinics[like MinuteClinic] — especially when they’re not supervised by a physician." [note that the podcast was in 2018, and we've seen many retail health clinics pop up since then]. I appreciate if somebody on the "other side" supply objective analysis to push back on this line of "liberal" analysis as well to make sure I'm not locked into some tendency.