kalap_ur

How do you diversify now? I presume you don't refer to stock portfolio, do you?

Yeah. Hyperscalers who are building compute capacities became asset heavy industries. Today's Google, MSFT, META are completely different than 10 years ago and market has not repriced that yet. These are no longer asset light businesses.

I wonder if there is somebody here high up in the MSFT stack who understands the tech/code but also oversees more stuff to be able to opine.

great perspective and wisdom nuggets.

Well, this sounds like a "no shit Sherlock" statement: >>Finding 3: Natural "overthinking" increases incoherence more than reasoning budgets reduce it We find that when models spontaneously reason longer on a problem (compared to their median), incoherence spikes dramatically. Meanwhile, deliberately increasing reasoning budgets through API settings provides only modest coherence improvements. The natural variation dominates.<<

Language models are probabilistic and not deterministic. Therefore incoherence _by definition_ increases as a response becomes lengthier. This is not true for humans, who tend to act/communicate deterministically. If I ask the human, to read a pdf and ask, is there a word of "paperclip" in the pdf? The human deterministically will provide a yes/no answer and no matter how many times we repeat the process, they will provide the same answer consistently (not due to autocorrelation, because this can be done across different humans). LMs will have a probabilistic response - dependent on the training itself: with a very well trained model we can get a 99% probabilistic outcome, which means out of 100 simulations, it will give you 1 time the wrong answer. We have no clue about the "probablistic" component for LMs, however, simulations could be done to research this. Also, I would be very curious about autocorrelation in models. If a human did a task and came to a conclusion "yes", then he will always respond with increasing amount of eyerolling to the same task: "yes".

Also, imagine the question: "is the sky blue?" answer1: "Yes." This has 0 incoherence. answer2: "Yes, but sometimes it looks like black, sometimes blue." While this answer seemingly has 0 incoherence, the probability of increased incoherence is larger than 0 given that answer generation itself is probabilistic. Answer generation by humans is not probabilistic.

Therefore, probability driven LMs (all LMs today are probability driven) will always exhibit higher incoherence than humans.

I wonder if anybody would disagree with the above.

It is not the scale that matters here, in your example, but intent. With 1 joint, you want to smoke yourself. With 400, you very possibly want to sell it to others. Scale in itself doesnt matter, scale matters only as to the extent it changes what your intention may be.

If there is one exact sentence taken out of the book and not referenced in quotes and exact source, that triggers copyright laws. So model doesnt have to reproduce the entire book, it only required to reproduce one specific sentence (which may be a characteristic sentence to that author or to that book).

You can only read the book, if you purchased it. Even if you dont have the intent to reproduce it, you must purchase it. So, I guess NVDA should just purchase all those books, no?

I paid $150 for a 64GB DDR5 in Jan 2025. That is today $830 representing 5.5x.

I think this analysis has little to say. What would be important to know how those $ are being spent, not where they are collected from. We do not know how those $ are being spent.

1. Doctors, Nurses, Administration (management and field administration), other. We need to know total employment and total salaries (including private practices).

2. OTC, prescription and hospital administered drugs (separated for acute, such as ER, and chronic, such as inpatient and elective surgery). We need to know how much is being spent on these, which is _potentially_ one of the culprits of large discrepancy between US healthcare vs European healthcare. What would be great to have these by large cohorts of population (<20; 20-65; 66-85; 85<) and maybe the top 5 buckets (i am guessing: cardiovascular - chronic; diabetes; accidents; hospice; dialysis)

3. Facility expenses (rent, maintenance, utilities, other contractor)

4. Other

Without these, very hard to opine reasonably on the state of affairs. And to be fair, I suspect there is a reason why proper expense breakdowns are not available.

It claims that they can print 12nm features with their particle accelerator. This looks weird.

There is no need. The personnel cost multiplier is 3.6x, the non-acute drug multiplier is 6.25x and the "something else" multiplier 77x. So we can debate that there are ~4x more people in the US and CoL is higher maybe by 0.5x turns, but there is no debate that the 77x shows something is awfully wrong.

My bet is the private insurance because we dont have transparent data on how much does the same procedure (broken down by personnel (doctor, nurse, admin), implied equipment amortization, rent, drugs) cost with and without insurance.

I could totally imagine that there are plus personnel expenses buried within the "something else", or acute drug prices, which are administered during an emergency at a hospital. But we don't know, because healthcare spending is a black box in the US.

oh wow, i sense a bojler elado, here.