balarjin
> an Oracle facility designed, financed, and contracted around an older GPU
> generation can be overtaken before it is fully online or fully monetized.
Every user of Nvidia's gear has thought of this, and Nvidia provides a solution: Newer components are (usually) compatible with older racks, and Nvidia will buy back parts you replace with newer hardware.
For example, let's say you have an NV72 with Blackwell-ultra. Rubin comes out, and you decide you want to upgrade. There is a SKU of Rubin designed specifically to work in the racks shipped with Blackwell. Use it, and you only need to replace the compute board (not the power system, cooling, networking, etc.)
Nvidia will buy back the Blackwell boards you remove (and sell them to someone else, with a warrantee).
If you want to do an upgrade that increases power use, you may have to get a new rack. All of the parts are modular, and all parts can be sold back to Nvidia.
Nvidia does this because they have to. Without it, many customers would consider other vendors when upgrading, or delay purchases in case something better comes along later.
This is not a new problem. The internet has been full of nonsense since the 1990s (at least). Gossip and malicious rumors are as old as language.
It is not realistic to stop the generation of nonsense (of the malicious or the stupid sort). We should train people to detect and ignore nonsense when they hear it.
If an oil company creates new technology to extract oil, then they are (obviously) a technology company. If an oil company outsources all technology development, they are not.
There is no "viable third platform" because people don't vote against the two existing ones with their wallets.
Can you name an example where a government succeeded in breaking up a monopoly, and the result was better for consumers? AT&T, IBM, and Microsoft were all "broken up", but it didn't make a difference for consumers. In every case, a competitor forced them to change their ways.
I don't like (most of) the content on those sites. This is not surprising. The content is aimed at people who can't use Signal. Such a person is different from me in many ways.
1. Too stupid to make a plan that could possibly work
2. Too lazy to execute on a plan that could possibly work
Knowing this is a useful signal. It tells me their judgment is terrible, and I should not listen to them ;)
Illegal immigrants are a source of cheap labor. Business interests tell politicians that without a pool of cheap workers, costs will go up, voters will feel inflation, and the politicians will get the blame.
The overall cost is not low. Some illegal immigrants get benefits. Some illegal immigrants commit crimes. Almost all illegal immigrants will use the healthcare system eventually. But those costs are not seen by the businesses that push illegal immigration.
"Voters perceive problems. As is common in a democracy, voters elect people who propose solutions. Why is no other political party proposing solutions voters see as worth voting for?"
If you don't think the problems are real, make the case to the voters.
As far as I can tell, other political parties have no better answer than name-calling. Am I missing some politician/party that is actually making a real argument?
In normal English, it would be from the average (median) voter. A party is "far-right" (or "far-left") if it is substantially to the right (left) of the average voter.
Assume that people vote for the person/party closest to their own view on an issue. 44% of voters choose the far-right candidate, meaning 44% of voters are closer to the far-right candidate than any other option.
How can this happen? Several ways, but only one is realistic: the other candidates are even farther away (to the left or right) from the position of the median voter. 40%+ of voters are themselves "far-right".
> The Nazi party got 37% of votes nationwide (not just in one state).
The Nazi party was not that far to the right of the median German voter at the time.
The far-left is using scary words ("Nazi", "Far-right", etc.) to tar a group they dislike (see what I did there?) They are counting on the fact that most voters can be influenced by group-think and name-calling. They are probably correct. But this is Hacker News, and the people here ought to be smart enough to see this for the stupid political smear that it obviously is. If you don't like a party, make the case that their policy is wrong. If vague analogy to the Nazis is the best argument you have, then you have no argument at all.
That is an option. It is usually not the best option.
Dividends are taxed. You are forcing shareholders to pay a good chunk of money in taxes.
Suppose Nvidia decides not to hedge the risk that competing products eat into their margins. Some years they get lucky and are flush with revenue. Other years they have much lower revenue.
Even in bad years, they want to keep paying to make future products. Cutting employee pay or mass firings would break their engineering org. To avoid that, they would need enough cash on hand to survive several bad years. Now they can't pay dividends because they need this extra buffer of cash! Hedging is a better deal for everyone.
> The shareholders are free to invest in competing businesses.
OpenAI is not (yet) public. Most Nvidia shareholders can't choose to hedge risk by buying shares in OpenAI, or any similar company (they are all private or a small part of a bigger public business).
IBM had a system of "blue dollars". If the server group sells a computer, it gets revenue ("green dollars", i.e. real money). The leaders of the server group are credited some fraction of the real money as "blue dollars", which they can give to teams that enabled their success. For example, if sales are being won based on a new feature of the XLC compiler, the server group leads would recognize them by sending their blue dollars to the XLC compiler team. Teams choose what to do (and whose problems to solve) to maximize the amount of blue dollars they can get. Promotions and pay go to people who drive blue dollar "revenue" for their team.
Why does IBM work this way? It is too big to be centrally planned: If the compiler team does amazing work, and the engineers get no credit, all the smart and hardworking engineers leave. The central planners in Armonk re-invented a market economy to figure out what work matters and reward it.
Silicon Valley is not centrally planned. The people at Microsoft did not go to a committee and ask the people who ran Sun Microsystems if it would be okay to release Windows NT. They mailed the CEO of Sun a coffin (1).
1. https://www.pcgamer.com/software/windows/the-team-behind-win...
If you can see the future, by all means invest in the one stock that will go up the most. If you can't see the future, diversifying into assets that have negative correlation (one goes up if the other goes down) lowers volatility at the cost of limiting possible upside.
Nvidia has a massive pile of money. Where should they invest? If they believe in what OpenAI is doing, investing in it makes sense no matter what hardware OpenAI chooses. If OpenAI manages to make something much better than Nvidia hardware, Nvidia's sales will go down as the value of OpenAI goes (way) up. If OpenAI fails to make something as good as Nvidia hardware, they buy Nvidia hardware.
It is odd how internet commentators seem to think that companies make investment decisions as a way to root for a team. They do not. That is not how finance professionals think!
If Xi had said "the CCP is illegitimate unless it takes Taiwan by 2025", he would be at real risk of being replaced in 2026. Anyone who opposed him (or just wanted power) would have a perfect reason to replace him.
When Xi makes a goal with a "deadline" he won't live to see, he takes no risk of personally failing to meet the goal. If his successor misses the deadline, he can say things changed, Xi failed to consider factors X,Y,Z, etc.
Spell out the details: Exactly who will not be permitted to speak under your definition of "freedom"?
If you say that only people with a billion dollars can be punished for speaking: Do you think the limit won't be lowered over time?
If the issue is that no person should be able to spend more than N million dollars in one "conversation", how do you allocate broadcast licenses when the price in an auction would be more than N million?
> Honestly, you surprised me with this statement alone. interested in learning? > do you think Trump wants to learn anything from AIPAC? > or Pharma lobby or Military Industrial complex?
Trump is unusually sure of himself. But no man rules alone: There is a massive org under every president trying to run a government. I know for a fact that people two levels of management under Trump (i.e. Trump is their manager's manager) do take meetings with people who lobby on the issue of health care policy and (for and against) military procurement. I would be shocked if people who are for/against working with Isreal are never in a room with anyone who can set policy.
Imagine that it is your job to come up with policy for a bug expensive government program: The VA hospital system, a military procurement program that buys airplanes, the part of the state department focused on MENA, etc. If someone asks to meet, claiming that they did academic/technical work which shows you can get a better result with a policy change, you will take that meeting. Examples:
1. An MD researcher shows you can cut drug costs by switching one medication for another.
2. An engineer has reason to believe current stealth coatings on fighter planes won't work against China's missiles in 5 years.
3. An economist thinks that an ally (Isreal, Korea, etc.) have demographics that will stop them from being able to field an army within 20 years.
Most people in government want to solve problems (and get credit for doing so). Most problems governments face have hard tradeoffs that voters don't understand/care about. Voters want unlimited healthcare at no cost to themselves, 100% success thwarting terrorists with no spying on citizens, military supremacy without paying much for weapons, etc. People asked to solve these problems are DESPERATE for new ideas to allow voters to see that they are getting good outcomes. Telling them you have a way around a hard tradeoff is music to their ears.
Different people (including politicians) have different priorities, assumptions, biases, etc. The argument that convinces one person won't land with someone else. The true value of a lobbyist is her ability to understand what a politician values, and present the argument that is convincing to him.
As an example: I met an EU citizen who thought fishing with hooks was immoral. It seemed cruel to her to put a hook in the fish's face. She traveled to the US to ask senators to ban it. Amazingly, she went to a few effective altruist meetups, networked her way into contacting the staff of two US senators, and managed to get meetings with both of them. But neither one introduced legislation for her.
What arguments could she have made?
1. The vegan argument: It is wrong to eat animals or make them suffer.
2. The environmental case: Humans should leave nature alone.
3. The public safety case: Some rivers have toxins that may accumulate in fish. Only a strict licensing system can keep people from fishing in polluted rivers, selling the fish to people who don't know the river had bad stuff in it, and getting sick.
<<I could come up with many more, but you get the idea...>>
"Senator X cares about the environment. Make the environmental case, and he is likely to listen to you. His constituents are not interested in veganism, and he wins elections by talking about standing up for the interests of ranchers. The vegan argument will not work, do not bring it up!"
Tailoring arguments to your audience can be dishonest if you take it too far. Not tailoring your argument at all is certain to fail: US Senators are busy. Handing them a 60 item list of all possible arguments, some of which they think are crazy, will cause them to ignore you.