US Political Discussion: Biden/Harris Edition (Rules in OP)

[Drew]

Laughable analysis and at no point have i said no to action on climate change. Just that the money has been spent in the wrong solutions.

By your logic you support the EU spending $254 Billion in an ETS that exempts the majority of heavy industries cutting a moderate amount of CO2 vs the US where they used direct action replacing coal with gas and saved far more Co2 with far less disruption and solidified their grid.

How does a business reduce their Co2 without changing their supply? Who controls supply? Government. The tail can't wag the dog. The american approach was right.

Is Solar right for Germany? No, they spent 4x more to get the same amount of solar power produced than America. That solution is not good.

Germany burns wood for fuel, which they get from the united states instead of clean nuclear. Is that right?

Europe Rethinks Its Reliance on Burning Wood for Electricity (Published 2022)

A new proposal would significantly rewrite E.U. rules on renewable energy, ending subsidies for biomass like wood pellets.

www.nytimes.com

It's a rich tapestry.

So, what, pray, was the correct amount of spending, if you think we "over-spent" to close the hole in the ozone layer? I'm not even sure why you think my analysis was "laughable," considering it was verbatim quote of your own post? Unless you instead meant the article we were discussing was "laughable," which makes me wonder why you posted it in the first place to try to support your belief that the hole in the ozone layer wasn't caused by CFCs? Which even the author of the study in question doesn't dispute?

 

[PuckishGuitar]

https://research.noaa.gov/2023/04/0...ting-chemicals-why-are-some-still-increasing/

Couple of standout comments from the same author about the study you are referencing.

The new study focused on global emissions of five CFCs with few or no known current uses — CFC-13, CFC-112a, CFC-113a, CFC-114a, and CFC-115 — chemicals with atmospheric lifetimes ranging from 52-640 years. In terms of their impact on ozone, emissions of these five CFCs combined in 2020 were equivalent to around one tenth of emissions from CFC-11, one of the primary ozone depleters in the CFC family.

While the CFCs presented in your chart are indeed rising, they are a fraction of CFC-11's effects which has been on a decline due to the Montreal Protocol (from over 400 at the peak to currently under 40 Gg/yr), with a spike in the latter 2010's before dropping dramatically in 2019 to below 2013 levels based on observations and attributed to new production in China. Note the rise in 2010 when Montreal went into effect and new HFCs started being manufactured, of which the five CFCs shown are byproducts of creation. CFC-11 is just one of several banned CFCs with outsized impacts that have had substantial declines - lo and behold the ozone is repairing now that total ozone chemicals in the atmosphere have declined significantly.

If emissions of these five CFCs continue to rise, their impact may negate some of the ozone recovery benefits gained under the Montreal Protocol, researchers said. The study noted properly destroying any co-produced CFCs would eliminate both any future impact to the ozone layer and a source of heat-trapping greenhouse gases.

He's still sounding an alarm at the rise if left unchecked.

I get what you are saying, the correlation isn't entirely clear and new data challenges old conclusions, but it's there in some way. It's tough to manage risk and determine where the cutoff is to provide resources to mitigate. Maybe we are overspending on solutions, and could cut it by 20% to get the same result. Or maybe in 10 years new research shows that we actually didn't meet our goals and should have done more. I don't expect to get into a car accident and maybe I never do, but I still pay my auto insurance. You have to act on the data you have to the best of your abilities. Finding out that you aren't entirely correct or more data is needed to get to the level of precision you want isn't a reason to not try at all.

 

[narad]

Aha, I see you cannot see proportionality.

X + Y = Z

X = our cutting CFC contribution to ozone depletion
Y = Unknown natural process to repair ozone depletion
Z = Actual ozone impact

We didn't know what our impact would be on Z but we thought X would be large (humans bad) and Y would be small (not a natural process) so any reduction had to come from us.

Turns out X is like 40% and Y is like 60% so whatever we spent, it didn't have the expected result. At the outset we didn't have any physical evidence about the CFC mechanism (it was not even able to be simulated in the lab properly, neither is it today), just an equation on paper, but decided to spend money on it anyway.

Now we could have spent money on X and not knowing the natural process Y might have been negative anyway. So we could have spent all this money and the ozone layer just depleted through a normal process so that would have made people spend even more money on X to lower our contribution.

Cost effectiveness bases were here, once again a desktop analysis based on a desktop calculation:

K Cost-Effectiveness of Chlorofluorocarbon PhaseoutUnited States and Worldwide | Policy Implications of Greenhouse Warming: Mitigation, Adaptation, and the Science Base | The National Academies Press

Read chapter K Cost-Effectiveness of Chlorofluorocarbon PhaseoutUnited States and Worldwide: Global warming continues to gain importance on the internat...

nap.nationalacademies.org

What's laughable is the fact that so much was unknown but still spent the money anyway. Now we know it has been reduced and only a small contribution could EVER have come from us. This is despite CFC emissions increasing and a full ozone recovery happening in the next 50 years.

View attachment 125732

I hope you learnt something today.

Dunning-Kruger over 9000!!! You didn't even actually address the clear questions of his comparatively short post lol

 

[Drew]

Aha, I see you cannot see proportionality.

Might want to reread the section about this being a closely moderated forum and avoiding personal attacks, buddy.

The assumption you're making was whatever we spent to close the hole in the ozone layer was exactly the maximum amount we were willing to spend for the intended result, and that we were unwilling to spend a penny more. Can you back that up? If not, if we were willing to spend 4x what we did, in theory, but spent 25% of our theoretical threshold past which point we're "not getting our money's worth" on saving the world and whatnot, and still got 40% of the desired outcome from 25% of the spend, we actually outperformed a fair amount.

 

[TedEH]

I still just come back here every once in a while to skim what's going on, and we still haven't moved past "doing good things for the world is too expensive, so to hell with it". ¯\_(ツ)_/¯

 

[MaxOfMetal]

I still just come back here every once in a while to skim what's going on, and we still haven't moved past "doing good things for the world is too expensive, so to hell with it". ¯\_(ツ)_/¯

Who would have thought you'd need to explain the return on investment of the planet not dying.

 

[MFB]

Who would have thought you'd need to explain the return on investment of the planet not dying.

 

[Drew]

I still just come back here every once in a while to skim what's going on, and we still haven't moved past "doing good things for the world is too expensive, so to hell with it". ¯\_(ツ)_/¯

I mean, the point of capitalism, ultimately, is to expand the size of the proverbial pie so that everyone can have a bigger piece, and an improved quality of life, and not "shoot the moon! let's see how big we can make this fucker!" Using some of that pie to address the risks of climate change, most of which are ironically attributable to some of the very same industrialization that allowed us to raise our living standards as far as we have, seems to be a pretty reasonable and fair use of our gains, as a society, I'd think.

Capitalism only gets kind of shitty when we start forgetting it's a means to an end, and not an end unto itself. See, well, this entire discussion, on the cost/benefit analysis of keeping the planet habitable.

 

[TedEH]

I mean, the point of capitalism, ultimately, is to expand the size of the proverbial pie so that everyone can have a bigger piece

Some conversations I've had recently would suggest that some see it as something that doesn't have a "point" so much as it just "is what it is, because that's the way it is", and the way it's always been, so suck it up and walk to school 10 miles backwards in the snow uphill by pulling on your bootstraps or whatever other dumb expression you want to use to ultimately mean "oh yeah, well I struggled before you, so therefore to hell with anyone else".

 

[Randy]

They Plugged GPT-4 Into Minecraft—and Unearthed New Potential for AI

The bot plays the video game by tapping the text generator to pick up new skills, suggesting that the tech behind ChatGPT could automate many workplace tasks.

www.wired.com

 

[narad]

They Plugged GPT-4 Into Minecraft—and Unearthed New Potential for AI

The bot plays the video game by tapping the text generator to pick up new skills, suggesting that the tech behind ChatGPT could automate many workplace tasks.

www.wired.com

Not exactly new potential -- this has been a main thrust of research at pretty comparable levels the past 6 months. But yea, more evidence against the ~it's just a tool sort of argument as it applies to job replacement.

 

[Randy]

Not exactly new potential -- this has been a main thrust of research at pretty comparable levels the past 6 months. But yea, more evidence against the ~it's just a tool sort of argument as it applies to job replacement.

Yeah not a new use as much as an example of nuance and adaptation.

 

[TedEH]

A nuanced and complicated tool is still a tool. ¯\_(ツ)_/¯

I still maintain that the "intelligence" on display here is the clever ways GPT and other programs are put together to get the desired result. Yeah, that's still a threat to jobs that can be automated - but it's not like the machines are going to spontaneously decide we're all obsolete. Someone somewhere is going to have to decide that they want this software to perform particular tasks and go through the process to hooking that up - which would be happening whether any set of semantics declared it to be "intelligent" or not. One way or another, technology eventually takes over some jobs.

 

[narad]

A nuanced and complicated tool is still a tool. ¯\_(ツ)_/¯

I still maintain that the "intelligence" on display here is the clever ways GPT and other programs are put together to get the desired result. Yeah, that's still a threat to jobs that can be automated - but it's not like the machines are going to spontaneously decide we're all obsolete. Someone somewhere is going to have to decide that they want this software to perform particular tasks and go through the process to hooking that up - which would be happening whether any set of semantics declared it to be "intelligent" or not. One way or another, technology eventually takes over some jobs.

Yea, but look at how far you've stepped back. "Someone somewhere is going to have to decide X and hook it up"? Hooking it up is near trivial because the "API" is just text. And once it's hooked up it does that for forever, at all hours, and can also adapt to changes on the fly. That's not a very high burden. And yes, technology eventually takes over some jobs, but tautologies aside, it's very capable and poised to take over enough jobs that it'll be a problem for a lot of people.

Like when you step back to last year and think about the process involved in planning tasks for success in minecraft, and how that relates to implementing and training component models to do comparably well, it just went from a very time-consuming expert level task to a largely AI driven one, in which the experts here are experts, but they weren't functioning really in that capacity. Putting that stuff together is not the PhD level ML that it was last year.

 

[narad]

btw, the Nvidia Keynotes are sort of sketching out this picture of the future. I think it's pretty accessible:

 

[Drew]

Yeah not a new use as much as an example of nuance and adaptation.

Did it get any good? Only scanned the article but it basically was saying this approach was more efficient than other forms of AI, but it didn't really get into how it would stack up compared to a human player.

A nuanced and complicated tool is still a tool. ¯\_(ツ)_/¯

I still maintain that the "intelligence" on display here is the clever ways GPT and other programs are put together to get the desired result. Yeah, that's still a threat to jobs that can be automated - but it's not like the machines are going to spontaneously decide we're all obsolete. Someone somewhere is going to have to decide that they want this software to perform particular tasks and go through the process to hooking that up - which would be happening whether any set of semantics declared it to be "intelligent" or not. One way or another, technology eventually takes over some jobs.

I remember seeing a finance article about how ChatGPT could be used to automate certain amounts of investment commentary, and correctly predict the market reaction to a certain bit of news. That sounds very impressive, particularly the latter. But, it's a little bit of a misnomer; it's actually not all that hard to predict the market reaction, in a general ballpark, to Apple missing on both earnings per share and revenue, or Chevron cutting their dividend unexpectedly. Or, in more my neck of the woods, the interest rate reaction to the FOMC raising the Fed Funds Rate in their June meeting. The hard part isn't figuring out what happens next - it's predicting that Apple was going to have an earnings miss, or the Fed was going to raise rates when the market thought they weren't (basically the story of all last year). When AI starts to get good at that stuff, well, things will get interesting.

Until then, maybe you can use AI to make a more nuanced modeled prediction of what the price of Apple would correct to after that miss. There's still value in that, and particularly in options trading there can be lots of money to be made in the difference between dropping to $178 and dropping to $169, but AI being able to pre-empt these things (and certainly, for public companies, theres potentially some big data type potential to try to measure sales volumes before earnings, aat both the retail (output) and supplier (input) sides, so maybe it's possible) would have all sorts of fascinating implications for the efficient markets hypothesis...

 

[narad]

Did it get any good? Only scanned the article but it basically was saying this approach was more efficient than other forms of AI, but it didn't really get into how it would stack up compared to a human player.

The point of that system is not about "goodness" but about the use of a language model to dictate a general plan and execute it. You're basically moving from a problem specific model trained more ground-up in the game world to an AI that is more general purpose and brings a bunch of general commonsense problem solving (and in this case, likely a lot of reading about minecraft strategies) to set agent behavior in the situated domain.

I remember seeing a finance article about how ChatGPT could be used to automate certain amounts of investment commentary, and correctly predict the market reaction to a certain bit of news. That sounds very impressive, particularly the latter. But, it's a little bit of a misnomer; it's actually not all that hard to predict the market reaction, in a general ballpark, to Apple missing on both earnings per share and revenue, or Chevron cutting their dividend unexpectedly. Or, in more my neck of the woods, the interest rate reaction to the FOMC raising the Fed Funds Rate in their June meeting. The hard part isn't figuring out what happens next - it's predicting that Apple was going to have an earnings miss, or the Fed was going to raise rates when the market thought they weren't (basically the story of all last year). When AI starts to get good at that stuff, well, things will get interesting.

Until then, maybe you can use AI to make a more nuanced modeled prediction of what the price of Apple would correct to after that miss. There's still value in that, and particularly in options trading there can be lots of money to be made in the difference between dropping to $178 and dropping to $169, but AI being able to pre-empt these things (and certainly, for public companies, theres potentially some big data type potential to try to measure sales volumes before earnings, aat both the retail (output) and supplier (input) sides, so maybe it's possible) would have all sorts of fascinating implications for the efficient markets hypothesis...

I think finance is one of the least exciting areas to apply this sort of AI, since a great deal of the data is already numeric and there have been decades of well-funded attempts to do basic document processing and tie that info into the models. I'm sure there's potential to have models read a wide swath of knowledge and make some connections between related pieces in the global economy, but if it involves patterns humans aren't particularly good at recognizing, it's not likely on the chopping block.

 

[Mathemagician]

I still just come back here every once in a while to skim what's going on, and we still haven't moved past "doing good things for the world is too expensive, so to hell with it". ¯\_(ツ)_/¯

Lol same. I completely stopped checking this thread regularly. I now look at it every few weeks to months at most.


Today’s topic - collective bargaining:

Few years ago I was saying people needed to make more money on average, and that increase in pay will mathematically have to come at the expense of slightly less profits for the large publicly traded firms/the shareholders many people work at.

Because inflation happens annually no matter what, so people needed to be earning $20/hr to have a decent minimum wage.

Got called a “communist” lol, for saying that shareholders would need to accept slightly lower returns in order for people to make more. Even though labor is only one input into cost of goods sold, and corporate profits are (as always) at record highs.

Then some shared that they’d be upset that “others” might get more than them. But that was because they had important jobs and themselves were barely made $20-30/hr, and they couldn’t grasp that they were themselves underpaid and should be pushing near double that.

Now it’s a few years later and I still hope everyone got massive fucking raises/promotions, because companies have kept jacking up prices and the wages haven’t kept up for most.

https://www.salon.com/2017/04/28/la...ny-gives-pilots-and-flight-attendants-raises/

Meanwhile analysts say that it’s “frustrating” that a company is increase worker’s pay.

So while as much as 30% of the country wants to pretend they’re “above” fighting for higher pay and then take the side of publicly traded companies that would fire them in an instant - in the name of “capitalism”. People really need to take things like unionizing and protecting their rights to demand better.

Because the fairy tale of “working hard to outward my peers” translates to the hard worker making $16/hr, while the “lazy” employee makes $15.50. But they should both be making closer to high 20’s/low 30’s at a minimum.

^like many Starbucks that unionized and prompted Starbucks to raise wages to between $15-23/hr in a desperate bid to stave off unionization.



TL;DR - Unionization isn’t “cheating at capitalism” nor does it reward lazy people more than it helps the average and hard working. Companies cheat at everything to pay people less and cut corners and sell less food for the same price. Things like bargaining let people demand time off, healthcare, and better pay with strict rules on overtime pay and time off.

 

[TedEH]

Because the fairy tale of “working hard to outward my peers” translates to the hard worker making $16/hr, while the “lazy” employee makes $15.50. But they should both be making closer to high 20’s/low 30’s at a minimum.

It's been my experience that most of the people pushing hard for "hard work" as a solution to everything are the ones in the $15 camp, as if it's a self-defense mechanism, where most of the "lazy" jobs I'm aware of make 30+ and (maybe more importantly) don't want to jump off a bridge at the end of the day.

I keep hearing people imply (or sometimes say outright) that "value" comes from struggling. It's a load of nonsense. It's rhetoric to justify squeezing every last ounce of life out of a person who doesn't know any better. Everyone should be "lazy" (read: have the freedom to have a life outside of their employment).

 

[Drew]

The point of that system is not about "goodness" but about the use of a language model to dictate a general plan and execute it. You're basically moving from a problem specific model trained more ground-up in the game world to an AI that is more general purpose and brings a bunch of general commonsense problem solving (and in this case, likely a lot of reading about minecraft strategies) to set agent behavior in the situated domain.

I mean, maybe I'm thinking of this a little too simplistically, and to be fair this was me legitimately wondering, and not some sort of a backhanded critique.

But, I DO think it kind of matters. Like, AI being able to train itself how to play minecraft based on being fed a manual and a bunch of strategy guides is pretty cool, for sure. But, in the context of wondering about the risk o job destruction AI poses, I do think it kind of matters if the outcome was "AI was able to execute an identified strategy, but far more slowly and not as well as human characters could learn," vs "AI was able to rapidly deploy an existing strategy to the standard of an expert player," vs "AI synthesized inputs well enough to come up with a new strategy not yet implimented by human players, and is performing at a very high level at or surpassing the best human players."

It's sort of the difference between crude and subhuman but functional intellect, human-like intellect, and "oh shit, someone unplug Skynet now, before it's too late" intellect.

I think finance is one of the least exciting areas to apply this sort of AI, since a great deal of the data is already numeric and there have been decades of well-funded attempts to do basic document processing and tie that info into the models. I'm sure there's potential to have models read a wide swath of knowledge and make some connections between related pieces in the global economy, but if it involves patterns humans aren't particularly good at recognizing, it's not likely on the chopping block.

The data side of finance, absolutely. Some of the rote side of finance, though - reviewing M&A agreements for an IB firm, writing strategy descriptions for a prospectus, or writing a quarterly update on investment performance; potentially there are some big efficiency gains there, which would free human analysts from essentially generating boilerplate, to doing higher value add analysis. Like I said, describing how the market will respond to a given stimulus isn't that hard... but it would take me at least a half hour to draft a good 3-4 paragraph response to a client question on what we expect would happen if the Fed hiked in June, and what we think would be necessary to see to make them hike, including grabbing a few data points an then re-reading/reviewing my output. AI could theoretically do that in 30 seconds, and I could start right at the review.