What would be the point of releasing good numbers a year before the election and then forcing yourself to release worse numbers just a couple months before the election?
This observation should be really obvious. The people who think this is nefarious believe they are thinking critically. What they fail to realize is that they are skipping the part where they critically examine their own thought process.
Surely revising the numbers downward close to election when people are paying most attention to them is exactly the opposite thing you would want to do to get re-elected? I'd argue the motivation for malfeasance only makes sense the opposite -- BLS does not want the administration reelected and thus is releasing worse numbers close to the election.
The scandal of never having released the numbers and a future administration finding out and publicizing it is way worse than just admitting it up front and hoping no one cares. Mark my words, when it comes to media pundits, they'll still use the old numbers from the original publications, and everyone will just go along. That's how these things work.
One non-boogeywoman scenario is that they were hoping the numbers would reverse and look good --so they bet things would get better, but they didn't. It's a chance people will take, sometimes.
Well a simple (cynical) take is that the current administration is no longer up for re-election, they’ve been replaced by a new candidate and so now it’s okay to blame the current administration.
Harris is literally the VP and takes an active role in advising and perhaps even running the current administration. Yes, she's not president, but one would hope the second-in-command would be taking on some responsibility in the administration. Thus, she very much shares this administration's success and failures.
"By the current administration to get re-elected" how? Without more specifics, this is a statement you could write about literally anything the current administration is doing or not doing, so it has no explanatory power at all.
how would releasing this new info be helpful? or are you saying this latest release is not by the current administration but the previous inaccurate release is?
I'm saying the Biden administration is ultimately honest of their own accord, but they rush to publish numbers and the media rushes to publicize the original numbers, in such a way as to make people overestimate jobs created. I've talked with people who will still cite old jobs numbers as reasons that Harris should be re-elected.
The article doesn't suggest doing this would have helped the current administration at all. Considering this was always going to get corrected around this time, quite the opposite. Any objective look at this is going to suggest it doesn't help the current administration get elected. So if we are considering conspiracy theories...
Unless I'm mistaken the teams that share a desk are supposed to come in 2 days a week. For teams that don't do this desk sharing you have to come in 3 days a week. So there shouldn't be an issue of trying to have each person use the desk for more than half the week.
They did eventually state 2 days a week for cloud although it’s still completely schitzophrenic.
“You must all be in the office the majority of the time, the company can’t operate efficiently without this”
“Oh we made it impossible people to be in the office the majority of the time? Ok for the people who our policies make it impossible to come into the office the majority of the time, which we previously required alongside statements that we need to do this we are now making an exception to this since it’s not actually needed and we also made the policy impossible to obey. For those not doing desk sharing though we totally still need you to come in 3days a week minimum because the company can’t operate efficiently even though we just said those guys over there can come in 2 days a week. Thanks for understanding our clear management practices everyone!”
I work on a couple sets of rules at Google. I enjoy it. Though I agree people really can make balls of spaghetti with it. Everyone’s macros are terrible except for mine . The configuration system is where things can get really out of hand: transitions, selects, flags, etc.
The VCs hate that Google and others bid up salaries. It means their investments don't go as far. He'd love for Google to lay off a ton of employees so his portfolio companies could get significantly cheaper labor.
Homebuilders are hesitant to set up new projects as the new builds will be more expensive than the already elevated house prices- higher labour and materials costs higher interest payments for the home builders And finally More expensive mortgages for the end clients. This kills supply and it will send prices into the stratosphere.
Yes. Each query needs to generate enough revenue to pay for the cost of running the query, a proportional cost of the cost to train the original model, overhead, SG&A, etc. just to break even. Few have shown a plan to do that or explained in a defensible way how they’re going to get there.
A challenge at the moment is a lot of the AI movement is led by folks that are brilliant technologists but have little to no experience running viable businesses or building a viable business plan. That was clearly part of why OpenAI has its turmoil in that some where trying to be tech purists where others knew the whole AI space will implode if it’s not a viable business. To some degree that seems to behind a lot of the recent chaos inside these larger AI companies.
It's always fascinating to see. The business portions are much easier to solve for with an engineering mindset, but it seems to be a common issue that engineers never take it into account.
This is saying nothing about "technologists" (or as they're starting to become derided as "wordcels": people that communicate well, but cannot execute anything themselves).
It would be... not trivial, but straightforward to map out the finances on everything involved, and see if there is room from any standpoint (engineering, financial, product, etc.) to get queries to breakeven, or even profitable.
But at that point, I believe the answer will be "no, it's not possible at the moment." So it becomes a game of stalling, and burning more money until R&D finds something new that may change that answer (a big if).
Inference will continue to shift toward profitability, Google and the rest are going to choke the cost to gain traction but as TPU costs come down and GPU costs scale up and models become more efficient the cost to infer will drastically scale down into profitability.
Many here know I personally despise AI, but to take off my tech critic hat for a moment, if subscription models don't work out, we still ultimately do have a bunch of pretty powerful base models to build on. If you can push inference to the edge and run on local machines, you can make your users bear the cost.
I'm fairly sure this is why there's a rush to get AI related capabilities into processors for consumer devices, to offload that computing so they don't bear the ongoing cost and it'll probably be more responsive to the user.
reply