Possibly a most archetypal example of Goodhart's law:
"When a measure becomes a target, it ceases to be a good measure"
Parents find kids get ahead by not only good grades but also by winning in science fairs, sports, or disability claims, so they figure out ways to game those systems . . . and, their value as a measure is eliminated.
With few exceptions, these are now only a measure of how well their parents gamed the system.
Doesn't the alignment sort of depend on who is paying for all the tokens?
If Dave the developer is paying, Dave is incentivized to optimize token use along with Anthropic (for the different reasons mentioned).
If the Dave's employer, Earl, is paying and is mostly interested in getting Dave to work more, then what incentive does Dave have to minimize tokens? He's mostly incentivized by Earl to produce more code, and now also by Anthropic's accidentally variable-reward coding system, to code more... ?
And if the Big Tech companies think it is so important to get all those compute and/or memory chips sooner and in larger supply, it should be no problem at all for those Big Tech companies to pay for the costs and then have priority access to all (or their portion of) the output for the future years.
OTOH, if they are insisting on not investing their funds or stock, and it is simply pressure on TSMC to take on the risk, TSMC should be very wary of taking on risk for those players (unless TSMC sees another advantage of producing into a likely glut or supply canyon shortly after the new fabs come online).
It was NOT the party balloon, it was the CBP's [0] massively irresponsible use of the laser weapon as a response to a threat it failed to identify properly, inside commercial airspace with zero coordination with the FAA or other air operators.
Any significant competence or even communication would have turned this into a literal non-event.
But this so-called 'administration' values only loyalty to a leader and find competence an impediment to implementing authoritarianism. So, here we are, squandering a half-billion dollars on a horde of idiots with a laser and a party balloon.
[0a] "Defense Department has a working relationship with Homeland Security, where CBP is headquartered, that allows its personnel to use certain military equipment for its objectives, testing, evaluation and use along the southern border.". Seems the laser was on-loan from the DoD, and the yahoos at CBP just decided to go wild.
So, the 'administration' specifically demands pharmaceutical companies move production to the US.
Then, when they do, and invest "hundreds of millions of dollars running Phase 3 trials enrolling over 43,000 participants based on FDA guidance..." to ge a new product approved, they get zero consideration.
Why would the executives not look at this and say "why should we remain based in the US at all?", and just move headquarters to a more business-friendly country?
Sure, the US is a big market, but if one cannot get anything close to a fair treatment of your products, you have no ability to work in that market anyway.
We have a true kakistocracy, a government by the worst, most corrupt, and least qualified.
I have no stake in this FDA thing, but it's very amusing to read economic arguments of why FDA should approve something because a US company spent $x and should be entitled to get such thing approved...
Very interesting when people complain about "corruption" when money suddenly cannot buy favors.
It is a bit funny given the origins of the FDA. That said, there does need to be some legitimacy. The value of the FDA is that the rules are being applied in a fair and logical manner. This sort of unfairness because of clear bias against mRNA tech is what is gross.
By all means, if there's a good test that's lacking or clear problems that can be demonstrated I absolutely want the FDA to use it's power to stop a drug from coming to market or even pull and existing drug from the market.
The problem I have is that when the FDA acts in an illegitimate manner, it opens the doors to questions like "do we even need this institution" or "should the power of the FDA be reduced." Both of which are things that are both bad and dangerous.
The FDA doesn't need it's powers curbed, it needs competent leadership that will follow the data.
I am absolutely NOT making any economic argument for approval or that an investment beyond $x should somehow guarantee approval! They are all professionals and know the risks and how to manage them.
And NO this is not about "money suddenly cannot buy favors". This is about illegitimate denial of even a hearing because the person occupying the agency's chair promotes idiotic conspiracy theories in order to cultivate a particularly stupid slice of the voting base. The corruption here is political.
The ONLY argument I am making is:
1) The govt agency should AT LEAST give a fair review of all applications, especially when the same agency previously gave guidance, which the applicants followed, on how the studies should be run.
2) The govt agency should NOT be issuing rare "Refusal To File" denials based on an obvious and unsubstantiated bias against a particular technology (in this case mRNA) because the leader of the agency is a conspiracy theorist.
If you aren't going to read the article before commenting, at least read the comment before replying.
"As radio waves move through a space and interact with people, they create distinctive patterns that can be captured and analyzed. These patterns are comparable to images produced by cameras, but they are formed using radio signals rather than light. "
The concept sounds not unlike like the multispectral imaging produced by Geordi's visor in TNG.
Seems conceptually possible, but likely too much computing power and observing time (to build up and learn each individual's pattern in that part of the RF band), at least in current times.
I'm sure it could be developed to work in the field, but what is the use case where it pays off to make the silly-money investment to make it happen? Especially so when it's far easier to simply notice pings and get better data when approximately everyone always carries their mobile phone.
>>The machines I fell in love with became instruments of surveillance and extraction.
Surveillance and Extraction
"We were promised flying cars", and what we got was "investors" running the industry off the cliff into cheap ways to extract money from people instead of real innovation.
The environmental review WILL help if it is used to adjust the mining techniques so they don't destroy everything nearby to do the work, or even if it jist creates a reclamation / restoration plan (and yes, factor that into the price, it's trivial). Taking too long is a problem.
Then make laws and punish the people who break them. It doesn't do any good to litigate before the project has even started. DUI is a problem and you solve it by arresting drunk drivers, not making them fill out paperwork before they go to the bar.
Your proposal is to do nothing and then make sure the entire thing causes orders of magnitude greater costs and damage which can be irreparable for centuries.
That has ALREADY been tried, and it was an absolute disaster, killing people, wrecking lives, and wrecking vast areas of ecosystems including driving species to extinction. You clearly were not around when rivers literally caught fire or when pollution required entire areas of cities and towns to be evacuated and dug up (look up Superfund Sites), costing taxpayers hundreds of $Billions.
A billion dollar mining operation is not a quick trip to a bar, and it is not putting personal liberties at risk to require planning.
It is far better to PLAN ahead and AGREE on the requirements up front so the company and investors can make sound profit projections and the ecosystem is protected. It is far worse for everyone to let them cause irreparable damage then hit the company/investors with crushing legal actions after the fact.
Yes, I agree that such reviews need to be expedited, the delay does no one any good. But doing the reviews is crucial.
Please read some history and lookup Chesterton's Fence before whinging about topics of which you are clearly ignorant
For vendors offering a valuable product that provides data or info, this will be a massive boost
For vendors whose offerings are primarily lock-in to a particular interface, ummm, good luck with that...
reply