nsx5200

IMHO, people are a lot more varied and complex than they can come up with labels for. People like to easily 'bin' people into 'normal' and 'abnormal' to make their lives easier. Most people think black and white, whereas some of us think in probabilities, with Gaussian curve dominating this space. If it's not covered by the 80% Gauss, they think that's 'abnormal', and would like to come up with a label for it, whether it's introversion, ADHD, autism, or whatever new fangled 'new' term they think they have 'recently discovered'. There's an incentive for the academics to try to come up with 'new' stuff, whether in nutrition, psychology, and sometime even in medicine in order to justify their own existence. Keep Gauss and the perverse of occasional academic incentive in mind, and it's much easier to differentiate true abnormality from just 'outside the norm'. Digesting other people's experiences(ex. Reddit, forums) also helps a lot. Again, just MHO.

A bit out of scope, but isn't one of Hyperliquid's end goal is to get everything financial onto their platform, implying that "everything financial" will be tokenized on their platform? I'm new to tokenization and Hyperliquid, so feel free to spank me. I can use an education. @thowed's article was very eye-opening, but unfortunately half of it was Greek to me.

The metric that I'd be really looking at for BTC is input cost(electricity, hardware, etc) per bitcoin/price of a bitcoin. When the input cost is greater than cost to mine(aka the incentive to keep the "most secure and powerful decentralized computer network" running), that network will have the incentive to now turn itself off. That means entities with large amounts of resources (e.g. North Korea, NSA, CIA) can control that network to siphon the BTC off inconspicuously while running trick play on their propaganda machines to mitigate any appearance of abnormalities until it can not longer be hidden. When that happens, BTC will be as worthless as any other cryptocurrency. The trends in AI compete with BTC in very similar type of resources, and if AI really gains widespread adoption, it may be the thing that actually kills off BTC. If I were holding a lot of BTC, I'd put a healthy dose of resources into AI just to hedge.

I'm nowhere near an expert in AI, but from my understanding, there are various ways malicious behavior can be embedded in these models, even if they're "open source". One of them is by training it to behave in an undesirable fashion. One such possible attack is essentially an embedded Manchurian Candidate with trigger word. Those behavior would be embedded in the weight themselves during training, and not through the policy code, so it would be not be easily visible by examination. Not saying the same thing can't be done with closed source frontier models provided by the US companies, but given the choices, IMHO, Chinese companies would prefer to use Chinese-based models whereas democratic countries would prefer to use models that originates from democratic countries. If anything, it would provide a cover for when the SHTF. This new AI world opens a whole new universe in the vector of attacks that we have not yet focused our attention to. That's how security normally evolves... first comes the race to add capabilities without little/few thoughts to security. Then as the capabilities mature and become more widespread, new attacks will force the attention to security. Also, it seems like those censorship are enforced in the policy code as well as in the weights. People have started modifying the weights to remove censorship... all black magic to me.

TBF, those Chinese open source models have been heavily modified to censor stuff, at best. Worse, may have hidden Trojan horse in them. We won't really know until researchers take some extensive poking to it. We're seeing Google release cheaper models to address the cost, and over time, one of the hyperscaler will provide a free/open source model to run on their infrastructure to sell the service. Different systems, different limitations.

In your case, if you can figure out a way to feed your investment history into AI, and have it generate it for you...

Unfortunately, this capitalistic system has a side effect of benefitting the rent-seeking landlords. It used to be literal land, but in a modern economy, equity in this type of economy is the new land. I admit it's an unwanted side-effect in the system and for those that do not already own enough of this new land to share in enough of its output, it is essentially a form of modern 'voluntary' slavery. In defense of the system, of all the systems that's tried, this is still the one that's shown to be the most productive. For those that realize how this system works can hack/escape it by practicing essentially what Buffet/Charlie have preached: spend less than you earn, and compound the savings (over long periods of time). For those that don't, the usual handcuffs to the system are done with tools of marketing, which many many people fall for. For those that can see past the handcuffs and practice what's preached, getting "rich" is more or less an inevitability. I agree that a pure investor, once getting past the point of FI(many of us here are also past the point of "early" in RE ), has a tough time justifying their value-add in this rent-seeking world. IMHO, it's a big driver for this Charlie quote: "I tried to be useful". How many of us here are actually pushing the envelope like Musk or the people he's inspired? All we can do is to try to push the our own envelope, however we can, and try to leave the world a bit better than before. In this knowledge-based society, one way is to continue to self-educate and try to leave behind some insights, not unlike what the ancient monks/nuns have done. IMHO, some "minor" value add as an active investor, as pointed out by Li Lu, is that active investors only get to keep and grow their wealth when they are right(*when done correctly, without corruption). So by nudging the price to where it should be, active investors are helping to decide where resources should optimally be directed in the economy a bit quicker. If that's still not enough, then I suspect given enough time, people will demand a new system, and we get to have another go at a new grand experiment. IMHO, capitalism is still an experiment, and we may be reaching the limits of it, at least in its current form.

AI is currently in the promotional phase, so a lot of it is given away for free, if not below cost. Just like how cloud/photo storage companies were giving away spaces with few conditions attached during the market share grab phase. After a while, the service providers will shift into profit/cost optimization, and free suddenly comes with a lot more conditions that steers people towards paid solutions. The fact that many of these AI/token providers have already started imposing some limits already, even on their paid subscriptions, portends to some serious cost issues. The primary lever for determining the eventual winner(s) of this AI war may be cost.

Thank you for sharing that video. At the end of the day, all this is "just money". Your view on life can change in an instant when something unexpected like that happens. All your life plan and what you value undergoes a dramatic reshuffling in these moments. When you're young, you feel like there's plenty of time, and so you trade time for money. When these moments inevitably arrive (hopefully through old age), your personal value of money depreciates precipitously relative to the stuff you really care about. IMHO, it's a worthy exercise to try to imagine such a life-changing event happening to you, and rearrange your life to mitigate/account for such an event. For all you know, these unexpected event can happen to you tomorrow. At the end of the day, money is just a construct/tool to motivate people to be productive, with plenty of side-effects, for better or worse. Don't let that be the main driver of your life. Instead, be a little selfish and think about what you would really like to do in life, and do enough of that as well. IMHO, there's a happy balance between YOLO and frugal FIRE. Cut out the useless stuff that marketing shoves down your throat, and you can achieve FIRE without much of the extreme frugality. Tagging @thepupil for this one.

It's a trade-off. Quality/time/money, pick two(Project Management Triangle). It seems like AI coding more or less still obeys that. For throw-away codes/short-lived codes, AI/vibe coding is probably fine. For longer living code, stability and proper design becomes more important. Like any tools and problems, things work better when they're matched up properly.

This is the same problem that teams have when deciding to off-shore development: spend time to formalized requirements/prompts, or spend time to do the actual design/development, with the hidden "requirements" communicated via human interactions(meetings/chats/"culture"/philosophy). With the off-shore teams, it's difficult to communicate culture and philosophy, so the end-product tend not to mesh as nicely. AI is an extreme form of off-shoring, with even fewer channels for communicating "culture"/philosophy correctly. I'm pretty sure this thread will explode if we start going down the rabbit holes of horror off-shore stories. If vibe coding works as marketed, then product managers would be writing the requirements for AI to code, but it seems like the reality is that people w/ formal training in software engineering is still needed to write those requirements for AI coding right now.

We're going to see a whole new suite of attacks and volatilities from these agentic trading. Who said value investing is dead?

If you view physical gold and BTC more or less equivalent, can you do pseudo pair trade on them and use that both of their volatility to ratchet up the growth?

It's that kind of nuance that non-technical people don't get easily. They would rather have a very simple narrative that AI will take over anything that requires "intelligence". Very black and white. For those non-technical project managers/finance analysts, a crash course in AI helps to tease out the possible from the (probably) impossible.

I like this part of the article: "Catherine Wu, head of product for Claude Code at Anthropic, says the visual flicker was a side effect of a software team moving at a rapid pace, [...]. The flicker has largely been fixed, she adds." The implication is that it is still not fixed, nor have they determined the root-caused of it to fix it. Probably because they asked AI to fix it, and AI mostly fixed it. Not saying it won't improve, but that's probably a good summary of the current state-of-the-art AI vibe coding. Seeing that AI is starting to break the open source model. I suspect GPL will come up with an update to address it, similar to how SaaS forced GPL to update to AGPL.