hydroreadsstuff

The landscape around sales seems to be changing. Specifically Square Enix and From Software are putting their games on sale extremely late or only lightly discounted. Final Fantasy 7 and 8 (21+ years old) are still 13€, Sekiro (1.5 Years old) is 40€. I waited 3 years for Dark Souls 3 (mainly because I didn't care enough) to be on sale for ~20€.

I remember I got Dark Souls 1 for 3€.

Just some anecdata. I think publishers are aware that their games are valueable longer longer, and they don't need those $5 buyers really. Perhaps this only applies to very popular games.

TLDR: I wouldn't bet on this being on sale for $5 ever

On the flipside: What else are you going to hold? How much are you allocating to cash, bonds, equity, real estate, gold, bitcoin(, collectibles, art, pokemon trading cards), etc? They all come with their own sets of risk. Think of it as another asset class, with different characteristics / (albeit changing) correlation.

We are at historically low interest rates. Equities are at historic P/E multiples in the U.S. and globally. Real Estate is getting more and more expensive compared to generated cashflow. Inflation/Deflation possibilities. Naturally, people are looking at other opportunities or hedges. Allocating portion according to your own situation and believe can make sense.

There also exists an argument on Gold vs BTC: - Gold is hard to transport - Same bar appears on multiple balance sheets - Paper gold ETFs - Price manipulation by central banks and hedge funds - Hardness / How much gold is added yearly - Ability to improve in use through software

Could be a very good collateral, compared to bonds which lose value through inflation.

It's really a personal decision. If you think your money should be elsewhere, then put it elsewhere.

Who would've thought that high-end chip manufacturing could be a (n enormous) bargaining chip.

I don't see the great keyboard either. My almost only reason to buy them.

Half the entries in your matrices need to be 0, then the Hardware will compress them and execute the matrix-matrix multiplication 2x as fast. On a tensor instruction level.

Semiaccurate and his twitter https://twitter.com/CDemerjian

The rumor mill suggets the volume is similar to Ice Lake (very very low)

This whole thread is why I always switch off auto-correct on any device. Mixing two languages is horrible. Unusual words / slang / abbreviations / street names / any unknown word throws it off. It's just painful to correct auto-correct, and the amount of facepalming involved.

Plus, I feel like my ability to type things correctly might suffer from crutches like this.

The same goes for voice recognition, when it tries to squeeze a specific name into a set of common words. It's just bizarre. In the past, when I tried the Google voice assistant or Siri, I always had to bring my best american accent (almost comical), otherwise they would often not understand.

The rate of new gold being dug up roughly matches global population increase. ~2% p.a.

Yes, this is indeed about the low bond yields paired with inflation expectation - low real yields.

But I would not hold my breath that this changes much after Corona. In 2019 the Fed tried quantitative tightening, which resulted in the December stock market correction, and a big overnight lending spike. From there they resumed QE (Which was one day supposed be a temporary measure in the great financial crisis). They did drop interest rates again in response to Corona, though. Not before.

J Powell already promissed rates would stay as low as they are for 2 years.

But right now the yields of the 10 year, e.g. are going below what a new issue should yield I believe.

Gold/Silver should lose some steam when the DXY goes up, or bond yields increase. (Both of which should also have a negative influence on stock and bond prices).

Also don't marry the idea of inflation to the CPI. There is asset inflation (think housing market and stock market), and inflation in certain products like food. But other things like transportation (perhaps of lower importance to poorer than for rich). It's not a black and white matter.

I like GDX as it contains both these and miners. FNV, WPM, SAND. There should be a lot more. The problem is that they are already extremely pricy. All gold related stocks had quite the ralley in 2020.

They lowered the rates for corona, but they expanded the balance sheet for that overnight lending rate spike before that.

From my point of view, as someone who is invested into an all-world blend like VT or VFFVX, it has been 2.5 years somewhat flat, and we still seem to be at a peak. https://www.multpl.com/shiller-pe .

I cannot really gauge risk in a systematic like you'd like to.

I'm generally pessimistic, because GDP growth is low, central banks have rates in a choke hold promoting misuse of capital. (In theory many existing companies shouldn't exist. Many are unprofitable, but they get by on cheap loans. Thereby taking away resources/people from potential/future profitable businesses). The ECB couldn't even raise rates in a supposedly economic growth period. The Fed did in 2019 while reducing the balance sheet, and it ended in a correction and lower rates and a restart of QE. Low rates also put a lot of pressure on banks and local banks. Economies and the world are increasingly financialized. The financial market adds little value to the real world aside from "productive" (think GDP-increasing / not consuming) lending. Consumer loans are awful. Buying your first house/land on a loan is good, though. Interest payments reduce economic output in the present. Individuals spend income of the future now, and depress future spending. E.g. interest payment of the U.S. are ~300bil/year right now ~10% of the budget. And for many economies and governments there is no end in sight. Ultimately this can only be solved by reducing spending (which doesn't seem likely for many governments), inflating it away, keeping rates at 0% forever effectively spending printed money, i.e. inflating it away. Also central banks can buy assets in the open market at face value to lower yields and swallow the coming defaults (which again means inflating it away, and also sending the clear signal that everyone will be bailout, perhaps again in the future). One aspect of this dynamic is currency demand through e.g. being an important world currency or having export surplus. Some countries can print money better than others.

When I say "inflate" I don't only mean according to CPI. If all the money goes into assets like equities or real estate, it is inflating that.

The stock market feels like a ponzi scheme more than anything.

I'm simplyfing here. Some countries are better than others. Perhaps a valid investment strategy would be to invest into indexes of good countries.

That felt a lot like rambling, but I hope you get some value from it. I'm interested in your thoughts.

Both of you make excellent points. I guess I'm jaded by the fact that VT (or e.g. VFFVX) has been almost flat for 2.5 years now. A pure US investment would have fared better.

I'm pessimistic about bonds because of the debt bubble (gov debt, MBSs, CLOs, corporate bonds) and low yields. Governments couldn't stomach higher rates, either. In the big sell-off we have seen in March everything went down together margin-call style. I guess it did lower volatility, but it also provided little upside.

Euro yields are even negative.

I don't see how any of the countries possible exit strategies could be good for bonds.

Thanks for the links!

Good point. I found this here: https://www.officialdata.org/us/stocks/s-p-500/1900 1929 - 1944 seems to be the worst.

He might mean debt monetization. When the Fed buys treasuries. I think the video didn't touch on that. M2 money supply changes nicely correlate with S&P500 performance. So there is truth to his statement.

I'd also throw out the assumption that the Fed will reduce the balance sheet eventually. They tried and failed in 2019. It's not going to happen. They are in fact "spending" (through the Treasury and their bond-buying programs), even though Powell likes to say that the Fed doesn't have spending power, but has lending powers.

The historic worst case for the inflation-adjusted S&P 500 is not being up from 1929 - 1987 - 58 years. https://www.macrotrends.net/2324/sp-500-historical-chart-dat...

Edit: This ignores dividend reinvestment as ummonk noted. https://www.officialdata.org/us/stocks/s-p-500/1900 suggests 1929 - 1944 was probably the longest time.

We are in unprecedented times. Highest debt, lowest bond yields, crazy P/Es, historic GDP drop (worst since 1929, or since WW2 for some). We don't yet know how inflation/deflation will play out. CPI is useless. A lot "inflation" is going into real estate / land and the stock market.

I don't think past wisdom applies anymore. If you put money into most products (including most index funds) on the stock market today you are in fact stock-picking / gambling.

Also tech companies are partially up for a good reason. Small caps have lost a lot earnings power. https://twitter.com/NorthmanTrader/status/128335311697343283... Much much more than tech stocks. E.g. tech stocks were almost the only ones doing any buybacks last quarter. Stay-at-home companies, solid/value companies, and cheap stocks is where hedgefunds seem to move to. But yes, some valuations are bonkers. EV stocks, CVNA, W, ...

Personally, I've been late to the party, and needed to learn on the fly, BUT: I now own a portion of gold and royalty and streaming companies (mining exposure with less risk of badly run mining companies. They lend money to miners in return for metal. I.e. miners don't need to worry about the price of the metal.) Mind that GLD and SLV are not fully backed by gold and silver. There is significant counter-party risk (E.g. if the bank goes bankrupt you might not receive anything). They also haven't really been audited, and JP Morgan employees have been charged for price manipulation of metals.

I reduced my bond exposure, and I will probably soon move into gov bonds exclusively. I don't want to gamble on MBSs, CLOs, corporate bonds being bailed out. (The Fed and ECB's policies skew the bond market. Ratings and yields are (increasingly) not great indicators of company health). There is little room for lower yields. Real yields (bond yield - inflation) will turn negative or are already. When rates eventually will be hiked, both stocks and bonds will fall. (See 2019 / December 2019)

I started stock picking. Mainly value, but really random stuff I find on Youtube investment channels and Reddit (Of course I read up on them, listen to financial calls. Balance sheet. Cash flow. Exposure to potential problems.). I sleep better at night with a cheaper stock that has less downside, even if it isn't riding the current trend/bubble.

Other than that I'd keep a good amount of cash for potential opportunities. In the current climate, stocks could fall any week, real estate will likely be on sale (depending on how many rich people will try scoop it up at the same time), or perhaps you can buy a share in a local business.

We are in unprecedented times (Lowest bond yields, highest debt, Crazy P/Es). Back-testing is only so useful. And only 20 years even less. We just had a GDP drop which is the largest since WW2 for some, or since 1929 for the U.S. Look at Japan's stock market index e.g. for what the future could look like.

Historically low bond yields: The room for rates to go lower is low. I would suggest to be careful with investing in them. If a rate hike comes eventually, both bonds (depending on maturity) and stocks (see december 2019 S&P 500) will fall. With negative real yields (bond yields - inflation), precious metals and related companies are lot more interesting. Perhaps REITs or some blend of solid dividen-paying companies. Anything that is somewhat inflation resistant.

Did you adjust your calculations for inflation? For the S&P 500, the absolute worst-case is 58 years of not being up, I believe. 1929 - 1987. Edit: This ignores dividend reinvestment. It seems to be 1929-1944.

If you factor out chip size and clock frequency, they do actually stagnate. They add new features, but legacy rendering performance is improved by single-digit percentage points, I believe.

If you look at A100 compared to V100 for e.g. FP32 FMA performance (not tensor). 14.1TFLOPS -> 19.5 = +38%, for 2x transistors (16->7nm), +35% SMs and 250W->400W is not that great. Note that NVIDIA uses boost clock for all A100 numbers and seems not have published any base clock so far. So their is a chance that actual sustained A100 performance is lower.

Turing GPUs have rather large dies. TU102 754nm vs GP102 471nm. So comparing them as is, isn't quite fair.

On the CPU front Intel used to use rather small dies for consumers (and even use die shrinks to just cram more chips onto a waver -> more $$$), but now that AMD forces their hand, they are giving in. But of course a lot this area goes into extra cores, not single threaded performance (diminishing returns there).

The main reason is the partnership with nvidia. You get nvlink to the CPU, which you don't get on Intel/AMD (I believe). Other than that, I don't think there is a real competitive reason. The support and future timeline from IBM is lacking. I'm honestly surprised they are still big friends.

The obvious answer is to test aggressively and quarantine the infected and direct contacts. Then the rest can be out and about.

This announcement is lacking a lot of specifics, which means it's worse than N1 / Ampere and more of the competition.

TNP always annoys me by its wordiness and lack of tables and numbers. It's incredibly low signal to noise compared to Ars and Anandtech.

The next episode 35 has some more infos on it, as well.

Quercetin is over-the-counter and has similar properties Studies according to the video below on Ebola and Zikka used 50-100mg/KG - a lot lot.

This Medcram video talks about Chloroquine and the others: https://www.youtube.com/watch?v=vE4_LsftNKM

Yeah that's right. up to ~700 Euros per month for the public health care system. If you earn ~60k or more, you can opt out and get a private plan for less money with other perks. But, overall German's pay ca. 7% (you) +7% (employer) into health care.

My guess is that Apple is getting rid of the charger rather than supporting USB, in favor of wireless charging.

I think NVIDIA is in a good spot with its strong software ecosystem, and an army of devtechs. Their architecture is pretty general, so that if the ML algorithms change they adapt better than say TPU. I can't find the video of the talk anymore, but late in the process of TPUv3 IIRC Google had to rebalance their hardware to account for new algorithm(change)s.

Second, most companies that want to deploy AI cannot afford to build their own custom hardware. Even most carmakers partner with Mobileye or NVIDIA.

These big companies also have research teams whose job it is to stay on the ball, research and develop AI techniques and influence how the hardware has to change.

As for eating your own dog food, I think NVIDIA does just that with their autonomous driving software stack, robotics kit, devtechs for customers and library optimization/porting.

With sufficient focus, funding and execution I think AMD and Intel can reach a similar spot.

That said, my hunch is that the legacy and compatibility that the general purpose hardware makers have to carry forward will become a problem. But until that happens (I.e. specialized HW delivers better bang/$) in a 2-10 years, they will likely figure out how to alleviate that and/or develop more specialized hardware.

Slightly off-topic: Is there data on how many people don’t put children into our world because of this?

Here is a recent training from AMD for Frontier: Link: https://www.exascaleproject.org/event/amd-gpuprogramming-hip... Video: https://youtu.be/3ZXbRJVvgJs Slides: https://www.exascaleproject.org/wp-content/uploads/2017/05/O...

It will be very interesting to see what AMD came up with to convince ORNL to switch from 7 years (or more?) of NVIDIA to something else. I don't think it's just a lower price. Perhaps AMD is doing a closer tie-in between CPU and GPU. Since they own both.

Right, I mixed up GPUs and CPUs, here.

Do ports (as in the picture) have independent pipelines, or do they execute certain pipeline stages of a big pipeline? I suppose, either way you can't issue to the same port in the same cycle.

This paper sheds some light on how instructions are divied up between units on NVIDIA GPUs. http://www.stuffedcow.net/files/gpuarch-ispass2010.pdf Table IV. Notice that fp32 mul is in the SFU and SP, while others are not.