Nice work! Would totally have used this when I was freelancing. Honestly love the serif'd fonts, would love to see everything serif'd tbh.
Also back when I had to do these (I used Wave) having a notes section was very useful to include a few things (i.e. I used to include conversion rates). Would probably be pretty easy.
The entire academic industry is in turmoil, the uncertainty on how bad things could get is probably the worst of it as Universities are having to plan for some pretty extreme outcomes even if unlikely.
For those who are questioning the validity of a 59% (or higher for some other institutions) overhead rate, your concerns are worth hearing and a review could be necessary, but oh my please not like this. This was an overnight (likely illegal!) change made with no warning and no consultation.
If the government decided that a cap was necessary it should be phased in to allow for insitutions to adjust the operational budgets gradually rather than this shock therapy that destroys lives and WASTES research money (as labs are potentially unable to staff their ongoing projects). A phased in approach would have nearly the same long-term budget implications.
Are there too many admin staff? Likely? Is this the right way to address that? Absolutely not.
For those who are unfamiliar with how career progress works in Academia, it is so competitive that even a year or two "break" in your career likely means you are forever unable to get a job. If you're on year 12 of an academic career, attempting to get your first job after your second (probably underpaid) postdoc and suddenly there are no jobs, you can't just wait it out. You are probably just done, and out of the market forever as you will lose your connections and have a gap in your CV which in this market is enough to disqualify you.
> For those who are questioning the validity of a 59% (or higher for some other institutions) overhead rate, your concerns are worth hearing and a review could be necessary, but oh my please not like this. This was an overnight (likely illegal!) change made with no warning and no consultation.
Why should the public believe that procedures that produced 59% overhead rates in the first place can be trusted to fix those overhead rates now? Sounds like a demand for an opportunity to derail needed reform by drowning it in red tape.
Also, what would be illegal about the change? Are the overhead rates in a statute somewhere? The grants certainly aren’t individually appropriated by Congress.
You seem to be unfamiliar with how indirect rates work.
First some basic math: if a project is budgeted at a direct cost of $500,000, the indirect rate of 60% applies to the $500,000, i.e. $300,000.
The total grant is thus $500K + $300K = $800K. The $300K indirect costs are thus 37.5% of the total. This is an upper limit, as many direct costs such as equipment do not get indirect rates applied to them.
Second, these rates are painstakingly negotiated with the NSF and NIH. Yearly audits to ensure compliance must be passed if funding is to continue.
Third, these indirect cost go towards to items such as electricity, heat, building maintenance, safety training and compliance, chemical disposal, and last by not least laboratory support services such as histology labs, proteomics core, compute infrastructure, and some full time staff scientific staff. Only a relatively small portion goes to administration.
Finally, scientists generally would welcome review and reform of indirect costs to ensure they get the maximize benefit from the indirect rates. However, DOGE is not interested in reform. They are interested in raze and burn destruction.
If DOGE gets its way, it will knock the Unites States off its perch as the world’s technological leader.
You can tell people the truth all day long. They don’t want to hear it. They’re convinced that academia is rotten to the core and none of your facts and figures will dissuade them.
For example I know at my institution every dollar, every piece of effort, is painstakingly tracked and attributed to funding sources. We have extensive internal checks to make sure we aren’t misusing funds. Audits happen at every major milestone. All of that effort is reported. It’s exhausting but the government requires it because we have to be good stewards of the funds we have been granted. No one believes it.
I’m not part of academia but was heavily involved in funding because of my position in student government while still in college.
While I won’t argue there isn’t waste (what endeavor doesn’t have waste?) it’s an incredibly tiny percentage (except in cases where there was actual fraud, which we also discovered and the Feds prosecuted and convicted people for).
The irony is that academia is so afraid of “waste” that I wouldn’t be surprised if colleges spend more money on the auditing and the compliances, etc than the actual waste they prevent.
I’ve had to deal with NIH audits up close. The amount of work devoted to compliance can make one question if the grant money is even worth it in the first place.
A big part of the reason indirect rates evolved is because the administrative burden to track direct costs is immense. How do you split up direct costs on an electric bill? Do you place a meter on each wall outlet and try to assign each amp to a specific job? Or safety training? Divide the safety meeting minutes by ….. ? It’s impossible. Which is why Vannevar Bush pioneered indirect costs. See the history section here:
> knock the Unites States off its perch as the world’s technological leader.
It's a funny thing. there is a distinct chauvinism to any citizen's nation. Every American is confident and absolutely positive that we are the best in so many categories. By what metrics? And who measures these? What about other nations who claim the top spot as well?
Before I travelled to Europe in 2008 I had some mental image of backwards, technologically inept populace that had old electronics and lagging standards and rickety brittle infrastructure. I mean you watch films and look at pictures and you see the roads and the old buildings and the funky cars and there's just a mix of things that are 500 years old or 1500 years back and thoroughly modern.
when I finally showed up in Spain I was completely disabused because all the electronics and the homes were totally modern and there were big box superstores that looked exactly like Target or safeway.
We went to shopping malls, watched normal first-run films in luxurious theaters that sold beer, and we rode around in cars/trains/boats, and I visited veterinarian and physician and hospital, and the medical treatment was indistinguishable from the American type.
I mean, this is one consumer's anecdata, but you've got to consider that we're ready to believe vague propaganda about #1 America First Outclassing The Solar System, and the fervent patriotism is perhaps not a 100% accurate lens.
Universities are designed to collect and disseminate knowledge worldwide. The top institutions and even the worst ones thrive on international collaboration. Think about how difficult it is to achieve and hold military superiority even. Schools are an effective equalizer, and globalist mindsets are the default.
the US is def not the best in many categories - though I suspect certain pockets of the US (overrepresented on HN) are like SV re: tech/quality of life and academia
many people I know - mostly [science/math/etc. denying] republicans think the US is the best at everything including healthcare (!!!) despite reams of data conclusively proving otherwise
my fingers are crossed that DOGE/Dump does something stupid enough to irritate the populace (and by extension a handful of senators/representatives to grow a mini-spine) enough to stop this destruction
Below the Ivy League and Premier type universities, many systems are based in/through a particular State, and so we could be more granular with a huge territory/populace and evaluate which States are ranked where for what types of research.
Further, it may be the case that Europe doesn't need/want a lot of high-tech, high-cost intellectual workers and opportunities that would drain brains from pools that do something more relevant, like soldiers, transport/shipping, or retail workers or HCPs.
in terms of scientific research though, America is ahead of much of Europe. It's historically been easier to get a good job in research in the US. Some research is also harder to carry out in Europe due to regulations. Now, whether the European lifestyle compares to the US is a different story. But when it comes to university-level research, it has been the case that there is just more money to throw toward it in the US, leading to more highly-cited papers. That might be changing, though.
> Also, what would be illegal about the change? Are the overhead rates in a statute somewhere? The grants certainly aren’t individually appropriated by Congress.
2024 appropriations (and it showed in many years before then-- Public Law 118-47. Statutes at Large 138 (2024): 677.
SEC. 224. In making Federal financial assistance, the provisions
relating to indirect costs in part 75 of title 45, Code of Federal
Regulations, including with respect to the approval of deviations
from negotiated rates, shall continue to apply to the National
Institutes of Health to the same extent and in the same manner
as such provisions were applied in the third quarter of fiscal year
2017. None of the funds appropriated in this or prior Acts or
otherwise made available to the Department of Health and Human
Services or to any department or agency may be used to develop
or implement a modified approach to such provisions, or to intentionally or substantially expand the fiscal effect of the approval
of such deviations from negotiated rates beyond the proportional
effect of such approvals in such quarter.
That says the indirects must be based on the existing regulations. The memo purports to rely on the existing regulations. It relies on 45 CFR §75.414(c)(1), which states:
> The negotiated rates must be accepted by all Federal awarding agencies. An HHS awarding agency may use a rate different from the negotiated rate for a class of Federal awards or a single Federal award only when required by Federal statute or regulation, or when approved by a Federal awarding agency head or delegate based on documented justification as described in paragraph (c)(3) of this section.
Subsection (c)(3), in turn, says:
> (3) The HHS awarding agency must implement, and make publicly available, the policies, procedures and general decision making criteria that their programs will follow to seek and justify deviations from negotiated rates.
Just based on a quick perusal it seems like the administration has a decent argument that the agency head can approve the 15% indirect by fiat as long as he or she comes up with a documented justification.
> So, an HHS division like NIH can use a different rate only for a “class” of grants or a “single” grant, and only with “documented justification.”
> There is nothing that says NIH could, in one fell swoop, overturn literally every negotiated rate agreement for 100% of all grants with all medical and academic institutions in the world, with the only justification being “foundations do it” rather than any costing principle whatsoever from the rest of Part 75 of 45 C.F.R.
Further, this doesn't allow a blanket adjustment to existing awards.
That is an argument in the opposite direction, but it overlooks two things.
1) The “documented justification” must reflect the requirements of subsection (c)(3), but that provision imposes no real substantive requirements. It’s a litigable, but the linked article concludes there must be more justification than the statute seems to require.
Note also that, amusingly, Kisor is still the law of the land and under that decision agencies still get deference in interpreting their own regulations.
2) The article frames the Congressional rider as prohibiting changes to the indirects. But the statute only prohibits changing the regulation, which HHS hasn’t done.
The statute just says the agency must use the existing regulations. The regulations were promulgated by the agency to govern its own discretion. The executive reads the regulation to constrain the civil service to a particular process, but allow the negotiated indirects to be overridden by the head of the agency with a documented reason.
You’re assuming that the regulation would constrain the head of the agency but why would that be the case?
Whether or not the head of agency is allowed to a drastic change like this doesn't change the fact that it is stupid. It's going to cost money in the long run.
At the *very* least you should be following the administrative rules act (requiring you to solicit 45 days for comments by effected parties) before making such a dramatic change in policy.
Courts absolutely love striking down EOs (of both Dems and Reps Admins) when they should have been following the administrative rules act.
You can file an APA lawsuit about anything. Nobody really calls APA violations “illegal.” It’s a “show your work” and “don’t be drunk or crazy” procedural law.
The “overhead” isn’t even overhead as most people understand it.
But the real question is why does the general public think 59% is too high? Irs an arbitrary number. Maybe an appropriate level of “overhead” is 1000%.
In reality the people who actually know anything about how this is calculated, across the board and across the political spectrum, do not think this is a major concern at all.
The only people who are complaining about it are the ones who hear the word overhead, have no concept of what it means other than taking a lay persons understanding that all overhead is unnecessary and are coming with the idea that anything above 0% is bad.
It seems like the better comparison from your article would be 1992, but really, having RFK Junior sitting there with a chainsaw is in no way comparable to 2013
It’s different because RFK with a chain saw might achieve change where Obama failed.
We have had 3 populist elections in the last 5 cycles. Obama 2008 was co-opted and Trump 2016 was stymied by Russia investigations. So this time there’s RFK and Elon and Tulsi with chain saws. If the people don’t like the results they can vote for Harris in 2028. But at least sometime tried to do what the winning party voted for.
These are cuts to enrich the extremely wealthy, not for a lean-mean-fighting industry. Your whole conception is off. They don’t need or care if the entire country does better overall, they care about personal wealth. It’s Obama wasn’t trying anything of the sort.
This is really it. Generally they gesture vaguely toward a notion of "administrative and bureaucratic overhead", without really understanding how that overhead actually cuts waste and improves research output by removing redundancies. If we were to zero out this administrative overhead, it would mean every professor would end up doing less research and more not-research.
1. Why should the public believe that they can fix it. Perhaps they can't, that's not entirely my point. My point is that if the government firmly believes that a change is necessary there are _simple_ ways of acheiving such a change without causing such chaos, waste, and hardship. Perhaps a phased in approach, or other mechanisms. Overnight shock therapy offers very little economic benefits while having very harsh personal and insitutional cost.
2. What is illegal about the change. The NIH overhead rate is actually negotiated directly between the institution and the NIH, following a process put into law. This is why a federal judge has blocked this order [1]. I'm far from a lawyer, but my read of this is that this is a change that would need to come through congress or a re-negotiation of the rates through the mandated process.
If they can’t be trusted to fix the problem themselves with a 5 year phase in period they most definitely can’t be trusted to fix the problem immediately…so I don’t get your point.
Everyone involved in the current process has an incentive to not change anything. If you go through the existing process with some five year target, the universities and bureaucrats will bleed you to death with procedures and lawsuits and lobbying, as they did with prior efforts under Obama. It’s the same way NIMBYs kill development projects. The only way to change it is shock and awe.
What article are we taking about? The response to “shock and awe” was rescind offers to students, not cut down on administrators or address inefficiencies.
That’s a temporary measure. The universities know that in he long run they need students but can cut administrators. But at least the immediate reaction is controlling costs rather than geering up to lobby and litigate their way out of it.
The US has a peculiar culture where elite academic institutions are very much willing to limit their numbers of students, so it's not clear to me that they will in the long run control costs. Large, prestigious US universities have historically preferred funding more administrators over more students.
Those elite universities are less like schools and more like towns, so the focus is not just on teaching students but on maintaining a community. Sometimes that means protecting the people you have at the expense of people you haven't met yet. In many cases, "more administrators" translates to "better town services", so it's nor surprising to me the preference to cut enrollment.
I’m not convinced that the rate, per se, is actually a problem. What is a problem is the structure. If a contract said “you get $1M to do X and your university gets $590k, paid pro rata by time until completion”, fine, and one could quibble about the rates.
Instead, the grant is for $1.59M, and each individual charge to the grant pays an extra 59% to the university, conditionally, depending on the type of charge and the unbelievably messed up rules set by the university in concert with the government. Buying a $4000 laptop? Probably costs your grant balance $6360. Buying a $5000 laptop? Probably costs $5000 becuase it’s “capital equipment” or “major equipment” and is thus exempt. Guess who deliberately wastes their own and this also the university’s and government’s money by deliberately buying unnecessarily expensive stuff? It gets extra fun when the same research group has grants from different sources with different overhead rates: costs are allocated based on whether they are exempt from overhead!
And cost-plus disease is in full effect, too. If the research group doesn’t use all their awarded money because the finish the project early or below estimated cost, the university doesn’t get paid their share of the unspent money. This likely contributes to grantees never wanting to leave money unspent.
Of course, DOGE isn’t trying to fix any of the above.
I'm curious about where you would draw the line on government workforce/spending reductions. What specific cost-cutting measures would go too far and make you withdraw support from Trump/MAGA-related initiatives?
For example:
- Complete elimination of federal workforce (RAGE)
- Full military withdrawal from NATO/Europe
- Dramatic cuts to essential services (eg, Social Security)
What potential actions would make you feel the downsides outweigh any benefits? I'm curious what your threshold is for acceptable vs. unacceptable changes.
>For those who are unfamiliar with how career progress works in Academia, it is so competitive that even a year or two "break" in your career likely means you are forever unable to get a job.
Honest question. If the job market is that competitive, why are we guiding people down this path that requires investing their entire young adult life? To me, it seems you've inadvertently made a case for cutting funding.
The big question is how should the government allocate the funding for basic research between career stages to maximize the benefit to the society.
If you focus on training PhDs, which is the American way, you get a steady stream of new people with fresh ideas. But then most PhDs must leave the academia after graduating.
If you focus on postdocs, you get more value from the PhDs you have trained. Most will still have to leave the academia, but it happens in a later career stage.
If you focus on long-term jobs, you have more experienced researchers working on longer-term projects. But then you are stuck with the people you chose before you had a good idea of their ability to contribute.
We aren't really. We are guiding people to get college degrees. However, undergraduate education and professional research are both done by the same institution. Further, that institution likes to have those professional and apprentice professional researchers work as teachers. The result of this is that undergraduates get a lot of exposure to professional Academia, so they naturally have a tendency to develop an interest in that profession. Given how small the profession actually is, even a small tendency here saturates the job market.
At this point, what profession isn't "small"? It feels like jobs are declining across all industries except for the most exploitative ones they can't easily outsource.
The people in charge don't want good action, they just want action and now. They want to damage these institutions. They have published and spoken extensively on this. That we keep letting their defenders change the narrative to pretend anything else and continue to give good faith WHEN THEY HAVE TOLD US THEY ARE NOT ACTING IN GOOD FAITH in insane to me.
BTW, interesting thinking on the action for action's sake governing style:
"The cult of action for action’s sake. “Action being beautiful in itself, it must be taken before, or without, any previous reflection. Thinking is a form of emasculation."
I have a PhD from a reputed US university and I agree with the fixed overhead aspect of this.
There is no reason students get a third of the grant money and live in poverty (30k per year) while the university hires a football coach for ten million and builds a new building every year.
This is exactly the way this has to be handled, the universities are intentionally making this look worse than it is for public sympathy.
Again please read my post carefully. There is a valid critique of overhead rates, but simply doing it suddenly in this manner has little added economic benefit in the long run, while ruining lives and creating waste/chaos in the short run.
You can make a strong argument these institutions require reform, but such reform should be done not overnight, and not through such broad strokes.
There is a very basis here for invoking “rule of law” where:
1) we’re talking about discretionary grants being made out of taxpayer dollars;
2) congress has delegated authority to make the grants and to the executive, including determining indirects; and
3) the executive action is being used to save money.
It’s also “the rule of law” in some sense when NIMBYs sue to keep a Ronald Mcdonald House from being built in their posh neighborhood, but that doesn’t mean we need to lionize it on that basis, or preemptively surrender to efforts to invoke the law to block reform. The universities can afford expensive lawyers with their 59% indirects, let those lawyers worry about it.
Universities have freedom in how to use grant money. The government had so far not bothered with controlling what they do with the money coming from the government. The situation is a bit like you donating to a charity and they spending it on executive bonuses.
Are you proposing that the government has to sign everything into law before taking any action? Can you think of why that might be a terrible idea?
Since when is it required that all the money in a budget be spent? Amounts are budgeted for a division and then it’s up to that division to operate within that budget. It doesn’t mean they have to spend every single dollar in the budget. In fact, it should be a goal to spend less than the money that’s allocated in the budget so that it can be applied to the next year. The idea that all the money has to be spent, regardless is part of the problem.
> Since when is it required that all the money in a budget be spent
Since Congress passed The Constitution's Appropriations Clause and the Congressional Budget and Impoundment Control Act (ICA) in 1974.
Because otherwise, the executive could just unilaterally shut down any part of the government at any time. Or siphon money from one department to another.
Which it can't by design, since congress controls the purse.
There's a process outlined in that act, under which the executive can report to Congress that it is reducing spending, and Congress must approve that reduction in order for it to take effect. That is the law of the land. The law is being broken.
If you don't like the budget, there is one governing body that can do something about it in this country. The legislature. They control spending, just like putting people in prison is controlled by the judiciary.
Strangely, all the people grousing about executive overreaches are dead silent on all this.
The sin has never been executive overreach, the sin was always an executive they did not control.
No. The law is not being broken. That act does not require the president or the administrative agencies to spend all of the amount budgeted.
There’s a process to handle scenarios, where the administrative branch feels that more money is not needed to be spent for the purposes that the money was initially allocated for. At the very least, there is a 45 day process starting from the point that it is determined to be a “deferral of the budget process” (continuous days that Congress is in session) that is allowed for Congress to pass a rescission bill. I don’t believe Trump’s been in office long enough for that process to even have taken place.
I think we are talking about different things here.
I am not writing in support of funding cuts.
I am strongly supportive of stopping universities from skimming most of the funding, and the research getting a tiny bit. Student researchers doing the actual work get less than minimum wage.
If you are surprised by the 'less than minimum wage' part, it's a bit of creative accounting by universities counting a 'tuition waiver' as part of your wages.
And yet everyone was arguing recently about how amazing Deepseek was because they operated on such a smaller budget and how the restriction of chips into China forced them to find an efficient solution to training an LLM model. Sudden and drastic changes don’t always result in bad outcomes; in fact, they can many times produce outcomes that were never possible without the shock to the system.
Most of the critics of the doge are arguing that the changes are too fast and that the system needs to gradually and systematically through a series of conferences and meetings come to a proposal that might be implemented sometime in the future.
Spurious. Football coaches are not paid by overhead dollars. but mainly by alumni that like football wins.
No, when a major for-profit company outsources research they pay way more than a 50% “markup”. Unless they go to a research university: then they pay much less, and just like the federal government they are getting a fine deal.
Yes, some rich foundations (Gates, Ellison etc) exploit the situation and do not pay full overhead costs: They are essentially mooching on the research institutions and the federal government.
So you don’t think that some of the money that gets sent to athletic directors to build fancy stadiums and pay for multimillion dollar coaches would’ve gone possibly to research facilities if those athletic departments didn’t exist?
No, I do not. Most health science centers do not have football teams ;-). I am at UTHSC in Memphis and I can assure you we do not send money to support the Vols in Knoxville. Worlds apart.
Really? Then why do they charge students athletic fees? Why do stadiums and athletic centers receive government grants and subsidies?
Here is a huge list of donations from various groups to an athletic endowment at the University of Alabama. Much of this money could have and would have been donated to schools directly for research and academic purposes instead the athletic foundations.
Alabama is one of the most successful Division 1 football programs in the nation. If these programs are so profitable, why do they need so much money for these endowments? And why all the money from governments and grants? Doesn’t add up.
> Here is a huge list of donations from various groups to an athletic endowment at the University of Alabama. Much of this money could have and would have been donated to schools directly for research and academic purposes instead the athletic foundations.
Because donations are made for tax purposes and virtue signaling … someone is going to get this money. Many of the donors are alumni and will donate money to the school. It was already targeted to the university athletic departments. It’s not a big stretch for it to be donated to another university department that has a direct academic role.
One of the options is “the donors kids” here. Donating money never leads to more savings than it costs, people donate because they support what they’re donating too. Sports boosters care about sports, not academics.
Doesn’t the football stuff fund itself through tickets, licensing, etc? It seems hard to believe research overhead grants are going to the football coach.
Only if the organization with the money wants to do that. Flip it around. Do you think the sports program at any major university pays for physics research facilities (or any topic outside of sports medicine)?
> For those who are questioning the validity of a 59% (or higher for some other institutions) overhead rate, your concerns are worth hearing and a review could be necessary, but oh my please not like this
This comment encapsulates a big part of why people like Trump. They are sick of inaction in the name of careful consideration and nuance.
And I kinda understand it. By analogy, I've see this many times at different companies I've worked at. Whether it's a 20-year scripting engine or a hastily put together build system that barely works but the entire project depends on, whenever someone suggests to replace it you just get reason after reason for why you have to move slowly and consider all the fine details. Months and years pass and the core system never improves because an "old guard" shuts down every attempt to change it. Finally, someone new comes in and calls bullshit and says enough's enough, and rallies a team to rebuild it. After a difficult process, you finally have something that works better than the old system ever could.
Yeah, I understand the story doesn't always end well, but the analogy above helps me understand Trump's appeal.
Let’s ask ourselves what happens when the story doesn’t end well and it’s a service that government has been providing. The answer may be lives are lost, the economy breaks, enemies win victories, etc.
Move fast and break things is fine in a competitive marketplace. It’s asinine for the government to do.
The answer is to elect better politicians who can nominate better heads and those department heads can drive the necessary change without succumbing to the old guard.
The federal workforce, as a percentage of all jobs in the U.S. was 4% in the 50's, decreased steadily to 2% in 2000 and has held roughly steady since then. (The source is https://usafacts.org/articles/how-many-people-work-for-the-f... second figure, and I'm taking total jobs as a proxy for the population that the workforce serves.)
The end of that period of reduction was Clinton's Presidency. Clinton's National Performance Review (NPR) started at the beginning his term in '93. It had goals very similar to the stated goals of this efficiency effort, but it was organized completely differently. He said, "I'll ask every member of our Cabinet to assign their best people to this project, managers, auditors, and frontline workers as well."
GPT4o: The NPR's initial report, released in September 1993, contained 384 recommendations focused on cutting red tape, empowering employees, and enhancing customer service. Implementation of these recommendations involved presidential directives, congressional actions, and agency-specific initiatives. Notably, the NPR led to the passage of the Government Performance and Results Act (GPRA) of 1993, which required federal agencies to develop strategic plans and measure performance outcomes. Additionally, the NPR contributed to a reduction of over 377,000 federal jobs during the 1990s, primarily through buyouts, early retirements, natural attrition and some layoffs (reductions-in-force or RIFs).
The recommendations that involved changes to law, the GPRA, were passed in both houses of Congress by unanimous voice vote.
I don't think the stated goals of the current efficiency drive are controversial. The problem is the method. I want to understand the basis for people supporting those methods, the "we've got to break some eggs" crowd, when the example of the NPR exists. In my opinion, it didn't cause conflicts between branches of government, didn't disrupt markets, and was wildly successful. It also caused much less disruption in people's lives, because the changes were implemented over several years with much more warning.
I, personally, don't think the real goals of this effort are the stated goals, but that's a different issue.
Elon proved with Twitter that large corporations can survive drastic, chaotic and insane levels of cuts/layoffs (80%+) and still survive. If DOGE waited to do things less drastically, nothing would ever get done. The cuts that are going through are nothing as drastic as what Twitter endured (except USAID) so I guess he is willing to risk short term disruption for long-term spending cuts and that the organization will reorganize and restabilize pretty quickly.
> can survive drastic, chaotic and insane levels of cuts
I get ads on X that are just videos of animals being slowly shot multiple times to death. There's also some for tools to slim jim car locks. None of the mainstream/normal accounts I used to follow (shout out SwiftOnSecurity) are there, and instead it's a hotbed of crypto scams and deranged vitriol. The site is still running, but is a shell of its former self, making so little money that Elon is trying to sue people (and now, abusing US govt payment systems) to force them to pay him for advertising.
I can see how if you think that's a success, that you would think similar actions with regard to the US government are successful. The necessary cuts he's making are not necessary, and I'm guessing you aren't impacted, so, given the general lack of theory of mind towards others, I'm not surprised you think they're ok. The rest of us out there who understand the idea of human suffering are concerned for our fellow citizens facing arbitrary and unnecessary pain as the result of a capricious court eunach's drug influenced decisions before the "restabilization" that will never happen.
This is false. Twitter is not the US government. And Twitter is certainly not the US scientific establishment which is dispersed broadly across the nation and which has taken decades to build up. Many research universities will shutter their research departments permanently if these overnight changes are implemented. This is especially true in smaller states like Alabama, which is why Republican Katie Britt is sounding the alarm. Moreover, many people will leave the field permanently.
Moreover, scientific R&D is a strange place to slash if cost savings are the goal. Medicare and Medicaid comprise over 50 times the NIH and NSF combined budget of approximately $50B. If we want to save costs, research into diseases like Alzheimer’s Disease is the way to go. Alzheimer’s currently costs the nation $412B per year [1], eight times the NIH annual budget. Therapies which delayed the onset of Alzheimer’s Disease by 20 years would nearly eliminate this cost.
Let’s be clear: DOGE is led by a self described autist who has little idea how government and broader society functions. The damage he will do if left unchecked is vast.
Twitter’s valuation has plummeted since Elon’s purchase.
And to the extent Twitter is still limping along it’s because Twitter due to its very nature benefits immensely from the stickiness of social networks.
For example, Facebook is almost completely junk. It hasn’t improved or been relevant in a long time. And yet it survives and makes tons of money simply because people don’t want to rebuild their networks.
There are many other examples where even minor cuts have been devastating. The classic example is of course GE, the ultimate example of cutting a company to the bone, which worked for a decade or so, but set the company up to essentially cease to exist after.
Then you have Boeing, a company in an industry with less competition than probably any other in the world and it’s struggling to make money because of this thinking.
Boeing and GE are inappropriate comparisons. Their cost-cutting maneuvers were primarily driven by moving existing, quality work to overseas contractors. It was simply about saving money without worrying about efficiency or long-term benefit. The overhead of managing contractors spread throughout the entire world is much more difficult than overseeing groups say within the Seattle Washington area. I really don’t see how this compares to reduction of work forces in government divisions. These government positions are not being moved overseas along with the complicated overhead of managing the groups all around the world.
The functions will end up being outsourced to contractors and bureaucracy will have to deal with managing them and their failures. This is exactly what has already happened to many departments and direct cuts to workforce will only worsen it.
Sorry, but I just don’t feel like you have the authority or knowledge to make that statement. How could you possibly know whether direct cuts will only worsen things? This is the type of issue that is argued between people inside of an organization that are fully aware of all of the factors at play.
It's not really a prediction I'm making, this already happened in the 70s with Nixon, the 80s with Reagan and the late 90s with Clinton. Direct cuts to employment in valuable functions have historically always ended up with core employees being replaced by armies of contractors which then need armies of bureaucrats to manage instead of doing things in house. It's why the US Digital Service started, for example. The issue has been argued for about 50 years now and the outcome has been pretty clear. It's the inevitable conclusion of firing federal employees but still wanting the program function to live on, you will inevitably end up with contractors and that has meant armies of bureaucrats to manage them.
This equivalence between a company that provides one app that, if it were to disappear, would hurt no one, and a government that has thousands of functions, many of which are life-and-death in both the short and long run, is just ridiculous.
Let's take one example. The Epidemic Intelligence Service (EIS) is a two-year post-residency program that trains health professionals in applied epidemiology. These officers are crucial for on-the-ground investigations of disease outbreaks. It's a 2-year program, with 50-60 doctors in each year. All of the first-year doctors in this year's program were fired by DOGE, so far, for a capacity reduction of 50%. Both years are in the 'probationary' civil servant category, so the jobs of the rest of them are still at risk.
I asked ChatGPT 4o for other examples, and it generated a list of 40. You can do that for yourself, if you're interested.
Twitter hardly ever made money before and after is in the same state now. Its contribution (anything?) to this country is far different than a government institution.
The comparison here isn’t encouraging and makes no sense.
Amazon made almost no profits for many many years others too. They follow a reinvest or expansion strategy and if investors believe it the stock goes up. It is not encouraging that Twitter lost 80% of its value under Musk's leadership and not something pne wants for the US Government which also does not work on a for profit basis. Ofcourse Musk fakes that he doesn't know that and promotes his unsubstantiated wins stories daily.
Amazon offered very obviously valuable and profitable services. I think we're starting to realize ad-based monetization is not how to maintain a billion dollar corporation anymore. I wouldn't have much aspects for Twitter even if Musk never took over. But he sure did accelerate things.
>ad-based monetization is not how to maintain a billion dollar corporation anymore.
Online ad revenue has been growing, 15% per year recently. Huge growth. That includes legacy networks like (decrepit) Facebook, which is seeing double digit growth, and the short form video frontier is growing considerably faster and constantly pushing out new ad/partnership models and is very much a strong growth industry in an of itself.
Ad revenue is more than sufficient to sustain a billion dollar corporation. It can and does sustain trillion dollar corporations, and the industry is currently in a strong growth phase with a lot of obvious green fields for innovation.
Twitter was an imperfect yet functional website before Elon. Elon fired most of the staff. Twitter then continued to be an imperfect yet functional website.
Hell, I remember ten years of HN saying "WTF does Twitter need so many people for??", and then those same people said "OMG Elon is insane to fire so many people!!".
Many people are more concerned about the messenger than the message. They’ll flip-flop their opinions solely based on who is doing the bidding.
A glaring recent example. If Biden had taken action like Trump has to negotiate with Russia to stop the Ukraine war, would the Democrats be screaming that Biden is a “Putin apologist”?
If Barrack Obama made statements about deporting undocumented immigrants (which he did), Democrats fall largely silent. If Trump makes similar statements, same Democrats scream fascism, racism, and Nazi/white supremacy.
Sure, rhetoric matters for style points, but the act supersedes stylish rhetoric. I’ll take proper action with clumsy rhetoric over inaction or improper action backed up with eloquent rhetoric, which is what most politicians provide.
This is an incorrect statement. Twitter’s revenue halved but its expenses were cut as well meaning its EBITDA doubled. The most likely conclusion on cash flow is that it went down actually, probably by a half in line with revenue (since revenue is a sign of flow in).
This is not the stunning retort to criticisms of Elon’s “fire them all” approach that some imagine it to be. It basically says “we cut expenses by 75% and only lost half our business.” Which half of the US government are you willing to lose, and are you sure you’re cutting the right 75% to lose the targeted half? Which half of the subjects that we fund R&D for are you willing to lose?
https://www.bankrate.com/investing/ebitda/ (“Some investors and analysts use EBITDA to assess the operating performance of a business or as a broad measure of its cash flow.”)
Increasing EBITDA by downscaling the business and severely cutting expenses is a common approach when turning around an unprofitable company.
https://altline.sobanco.com/ebitda-vs-cash-flow/ ("EBITDA and cash flow are both important financial metrics, but they serve different purposes and provide different insights into a company’s financial health.")
We can quote secondary sources back at each other all day, but it's somewhat pointless because the truth is what I said already: EBITDA and revenue are merely indicators for cash flow, not synonyms. You used the wrong words dude.
I also noticed you only replied on a pedantic point while leaving the substantive questions on which half of the government and research funding you'd like to see gone (and how these cuts target that half) as an exercise for the reader.
I think it’s common for people to refer to “cash flow” (without referring to OCF or FCF or whatever specifically ) when they mean EBITDA, but I’m happy to be wrong about that. I’m not a financial analyst. But as you acknowledge, EBITDA is an indicator of cash flow. Is there a difference between the two measures that you think is relevant to X? X is increasing how much money they’re making right?
I'm glad we agree that cash flow is not the same as EBITDA.
The question we are talking about is whether Twitter makes more money now versus before Musk's take over. If "makes more money" means revenue, then the answer is a definitive no, it does not make more money now. If "makes more money" means profit, then the answer is that we don't know but probably not because profit is found after ITDA (hence the B in EBITDA) and we know the ITDA is substantial for Twitter given how it was acquired.
So yes there is a difference between cash flow and EBITDA that is germane here, and the difference is that cash flow doesn't help us answer the question that we are asking while the one piece of information that we do have (revenue) tells us the opposite of the answer you're trying to imply.
Great euphemism for “In one month, we’re going to kill tens of millions by withholding food/medical care and permanently destroy institutions that took a century to build”.
This is the most infuriating part of this. Musk acted like a moron and overpaid for twitter. Then cash constrained, he rapidly cut things to save money. Now twitter is completely diminished in its reach, at an all time brand low, and at real risk from competitors.
Meanwhile, the companies Musk built that actually have dominated their space are big idea innovators like Tesla and SpaceX. Musk wasn’t successful because he’s a good penny pincher, he was successful by burning cash on big ideas and talented people.
But somehow we decided its case 1 that we’ll apply to the government.
Case 2 had a lot of safeguards around Musk to keep him isolated from the talented people. But Case 1 made Musk feel better. So we know which one he prefers. Not like he's going to suffer the losses the most.
It seems Twitter is in a death spiral. That is the model to apply to scientific research and academia that has powered Americas dominance for the past 100 years?
Sure, if we detonate all nukes, I imagine 20% of humanity will survive. "We" won't die out that easily. Me and you are probably dead, though. Statistically speaking.
Elon proved with Twitter that he doesn't know what he is doing. Huge loss, zero lessons. If US ends up being downsized financially and ethically the way Twitter has, that will also provide zero lessons for Musk.
Having spoken with people who worked there, Twitter built a system for which the technical its mostly ran without much help. So it’s not surprising that you can still tweet with most of the staff gone.
> Twitter that large corporations can survive drastic, chaotic and insane levels of cuts/layoffs (80%+) and still survive.
US Government is not Twitter, but yeah, I can see some element of it. Now I suddenly remember, the various comments here convinced me that X/Twitter will be dead in just two months. Yet, it's still around. Not that I care to make account there or bother looking at it, but I figure if it kept crashing, it would show up in the news. Maybe everyone moved to bluesky and it just doesn't have any customers?
So you're still just going to attack the most extreme interpretations and dismiss the truth because it's not literally dead?
Digg is still alive. Myspace is still alive. These big brands don't literally die often; someone will want to try and play around with it.
>but I figure if it kept crashing, it would show up in the news. Maybe everyone moved to bluesky and it just doesn't have any customers?
Twitter is technically stable in terms of servers. They did a good job doubling up by 1) minimizing the load needed by making users sign in to see more than a literal permalink (you can't even see comments anymore) while 2) being a bait to get more new accounts to report on engagement. Not that that matters now since Twitter is no longer publicly traded.
It's the everything else around it that caused it to plummet.
> So you're still just going to attack the most extreme interpretations and dismiss the truth because it's not literally dead?
I can attack my own interpretation, that seems fair, doesn't it? You can have your own interpretation, and then attack it, (or not attack it) later.
> Digg is still alive. Myspace is still alive. These big brands don't literally die often
I had no idea! That's cool. I never really used either one that much. I do know yahoo is still around. One difference is I periodically I see links to X/Twitter. I never see people link to digg or myspace. But sounds like you have a different perspective, which is also cool.
> dismiss the truth because it's not literally dead?
Just to be clear, the truth for me is my impression from reading HN about 2022 or so. Namely from comments like these:
> (Nov 18, 2022) I very much doubt that. Twitter must have had some bloat, but there's no way that 80% of the workforce was bloat. I'd be extremely surprised if Twitter(as in, the website/app, not the registered legal entity) still exists and works by the end of this year.
I agreed to them at the time. So not sure how "truth" and "dismissing" applies; it's really just an impression. Am I allowed to dismiss my own impression? Seems odd to object to that...
>I can attack my own interpretation, that seems fair, doesn't it? You can have your own interpretation, and then attack it, (or not attack it) later.
Your comment just gave some vibes that because Twitter didn't literally die 2 years ago (as you and others predicted) that it seems that introspection was completely proven wrong, "Yet, it's still around". I just simply wanted to assert that being nearly dead doesn't exactly inspire confidence, even though the doctor was technically disproven by his statement of "you'll be dead in 6 months".
There is also a good case to be made that the prices being bandied around are actually much too high [1]
TL;DR is three major factors:
1. The agencies that are doing the estimates are _very_ bad at exponential development curves (cough cough IEA estimating solar [2])
2. Unfortunately much of the developing world's economy is not growing as fast as we previously thought it would (similar thing happening with birthrates)
3. Many costs are absolute and _not_ marginal, which is just wrong IMO. We are going to need the energy either way, we should be talking about the "green premium" (as far as it exists), not how much it'll cost to generate XX TWH of energy
If you turn off your gas generator and replace it with solar + batteries, you will spend the entire cost of solar + batteries plus the decommissioning cost of gas (that may be negative if you can sell some parts) to go back to exactly the same point you were before.
So, no the cost is only marginal if you accept you will follow the depreciation curve of you infrastructure. And that's way too slow to reach the goal.
We have some crazy incentives to install new gas boilers in MA. I very specifically wanted to switch from gas to an air source heat pump and found there was only one company in the area who was willing to quote it (alongside their own, much cheaper, quote for a gas combi boiler) and their quote was outrageously non-competitive with local fossil fuel burning (in large part because our electricity is around $0.30/kWh, but also because they were the only supplier and even they didn't really want to do the work).
Even if the ASHP lasted forever, required no maintenance ever, and you had to buy a new gas boiler every 10 years, it would literally never make economic sense even if there weren't $2500 incentives on the gas boiler, but the movement on electric rates is definitely in the wrong direction if one wishes to displace natural gas with electricity (even at 400% efficiency).
Every year that things stay like this is pushing back the likely time to next re-evaluation for that property by another 20 years.
The price of home solar and batteries is dropping to the point that $0.30/kWh is becoming untenable in any home that has a decent amount of roof space. You’re better off financing a rooftop solar plant and buying 3-4 days of storage, even if you remain tied to the grid. Insofar as those costs are being driven by generation, the declining price of solar should eventually place an upper bound on what people will pay for electricity. Even if you don’t live in a sunny place and even if net metering pays $0, with a few days of storage you can reduce your grid consumption to the point where your actual need to consume expensive electricity becomes a tiny fraction of your overall usage. I think this will tend to push costs downwards.
Even for people who don’t have the space or capital to install their own solar, this will happen writ large as the US builds out utility scale solar, wind and storage.
Unfortunately, we have a 100 year old slate roof, which makes solar some mix of difficult, expensive, or not advised. At the exact moment of maximum heat demand (both seasonally and time of day), solar generation is at its lowest.
I do hope that slate lookalike solar tiles become advisable and cost-effective as I’d be happy to pay a small premium to generate and store locally.
I sympathize. We can't install solar either because of roof pitch and trees. I think my overall point is that if many customers can install solar and bypass expensive generation, this will tend to put a downwards pressure on generation prices in the long term.
Solar aside, I am thinking about installing a battery system that time-shifts from low-rate times to high-rate times. It's almost cost effective now.
>So, no the cost is only marginal if you accept you will follow the depreciation curve of you infrastructure. And that's way too slow to reach the goal.
The linked article also mentions a way less aggressive timeline, which means there's less of "tear out existing equipment and replace with renewable" going on, which raises costs. Moreover, the argument isn't that there's no such costs, only that they're being overestimated.
Quite a fun paper, but very difficult to draw conclusions from. Their headline finding is that they created a map between resting state and "watching a movie".
This is, for better or worse, the kind of research you can only do at institutions which have free fMRI scanning (MIT, Princeton, Harvard, etc.). No behavioural links, only very detailed activation maps that we can't really draw conclusions from. A missed opportunity IMO to spend these kinds of scanning resource on some kind of more narrowly focussed task with some behavioural outcome they can link to brain data.
To me the issue with AI generated code, and what is different than prior innovations in software development, is that it is the the wrong abstraction (or one could argue not even an abstraction anymore).
Most of SWE (and much of engineering in general) is built on abstractions -- I use a Numpy to do math for me, React to build a UI, or Moment to do date operations. All of these libraries offer abstractions that give me high leverage on a problem in a reliable way.
The issue with the current state of AI tools for code generation is that they don't offer a reliable abstraction, instead the abstraction is the prompt/context, and the reliability can vary quite a bit.
I would feel like one hand it tied behind my back without LLM tools (I use both co-pilot and Gemini daily), however the amount of code I allow these tools to write _for_ me is quite limited. I use these tools to automate small snippets (co-pilot) or help me ideate (Gemini). I wouldn't trust them to write more than a contained function as I don't trust that it'll do what I intend.
So while I think these tools are amazing for increasing productivity, I'm still skeptical of using them at scale to write reliable software, and I'm not sure if the path we are on with them is the right one to get there.
It isn't an abstraction. Not everything is an abstraction. There is a long history of tools which are not abstractions. Linters. Static code analysis. Debuggers. Profiling tools. Autocomplete. IDEs.
I can’t tell if this is an argument against the parent or just a semantic correction. Assuming the former, I’ll point out that every tool classification you’ve mentioned has expected correct and incorrect behavior, and LLM tools…don’t. When LLMs produce incorrect or unexpected results, the refrain is, inevitably, “LLMs just be that way sometimes.” Which doesn’t invalidate them as a tool, but they are in a class of their own in that regard.
Yeah, they are generally probabilistic. That has nothing to do with abstraction. There are good abstractions built on top of probabilistic concepts, like rngs, crypto libraries etc.
My wife studys people for living (experimental cognitive psychologist), the quality of MTurk is laughable, if that's our standard for higher level cognition then the bar is low. You'll see the most basic "attention check" questions ("answer option C if you read the question") be failed routinely, honestly at this point I think GPT4 would to a better job than most MTurkers at these tasks...
She has found that prolific is substantially better (you have to pay more for it as well), however that may only be because it's a higher cost/newer platform.
My take is the tasks on Turk are awful and will drive away anybody decent.
I had a time when I was running enough HITs to get a customer rep and felt I was getting OK results. I wanted to get better at running HITs so I thought I would “go native” as a Turk and try to make $50 or so but I could not find tasks to do that were at all reasonable. Instead they’d want me to “OCR” a receipt that was crumpled up and torn and unreadable in spots and said they’d punish me for any mistakes.
> In the first batch of participants collected via Amazon Mechanical Turk, each received 11 problems (this batch also only had two “minimal Problems,” as opposed to three such problems for everyone else). However, preliminary data examination showed that some participants did not fully follow the study instructions and had to be excluded (see Section 5.2).
If they stuck to the average Mechanical Turk worker instead of filtering for "Master Workers," the parent's conclusions likely would've aligned with those of the study. Unfortunately, it seems the authors threw out the only data that didn't support their hypothesis as GPT-4 did, in fact, outperform the median Mechanical Turk worker, particularly in terms of instruction following.
> Unfortunately, it seems the authors threw out the only data that didn't support their hypothesis as GPT-4 did, in fact, outperform the median Mechanical Turk worker, particularly in terms of instruction following.
MTurk, to first approximate, is a marketplace that pays people pennies to fill out web forms. The obvious thing happens. The median Mechanical Turk worker probably either isn't a human, isn't just a (single) human, and/or is a (single) human but is barely paying attention + possibly using macros. Or even just button mashing.
That was true even before GPT-2. Tricks like attention checks and task-specific subtle captcha checks have been around for almost as long as the platform itself. Vaguely psychometric tasks such as ARC are particularly difficult -- designing hardened MTurk protocols in that regime is a fucking nightmare.
The type of study that the authors ran is useful if your goal is to determine whether you should use outputs from a model or deal with MTurk. But results from study designs like the one in the paper rarely generalize beyond the exact type of HIT you're studying and the exact workers you finally identify. And even then you need constant vigilance.
I genuinely have no idea why academics use MTurk for these types of small experiments. For a study of this size, getting human participants that fit some criteria to show up at a physical lab space or login to a zoom call is easier and more robust than getting a sufficiently non-noisy sample from MTurk. The first derivative on your dataset size has to be like an order of magnitude higher than the overall size of the task they're doing for the time investment of hardening an MTurk HIT to even begin make sense.
This is just coming up with excuses for the MTurk workers. "they were barely paying attention", "they were button mashing", "they weren't a single human", etc.
It turns out that GPT-4 does not have those problems. The comparison in the paper is not really fair, since it does not compare average humans vs GPT-4, it compares "humans that did well at our task" vs GPT-4.
> This is just coming up with excuses for the MTurk workers
No. The authors are not trying to study MTurk market dynamics. They are trying to compare humans and LLMs.
Both questions are interesting and useful. This study is only asking about the second question. That's okay. Isolating specific questions and studying them without a bunch of confounds is one of the basic principles of experiment design. The experiment isn't intended to answer every question all at once. It's intended to answer one very specific question accurately.
LLMs can both be worse at Mensa tasks and also better than humans at a variety of reasoning tasks that have economic value. Or, LLMs can be worse at those reasoning tasks but still reasonably good enough and therefore better on a cost-adjusted basis. There's no contradiction there, and I don't think the authors have this confusion.
> The comparison in the paper is not really fair
The study is not trying to fairly compare these two methods of getting work done in general. It's trying to study whether LLMs have "abstraction abilities at humanlike levels", using Mensa puzzles as a proxy.
You can take issues with the goal of the study (like I do). But given that goal, the authors' protocols are completely reasonable as a minimal quality control.
Or, to put this another way: why would NOT filtering out clickbots and humans speedrunning surveys for $0.25/piece result in a more insightful study given the author's stated research question?
> It turns out that GPT-4 does not have those problems.
I think the authors would agree but also point out that these problems aren't the ones they are studying in this particular paper. They would probably suggest that this is interesting future work for themselves, or for labor economists, and that their results in this paper could be incorporated into that larger study (which would hopefully generalize beyond MTurk in particular, since MTUrk inter alia are such uniquely chaotic subsets of the labor market).
For me, the problems with the study are:
1. The question isn't particularly interesting because no one cares about Mensa tests. These problem sets make an implicit assumption that psychometric tools which have some amount of predictive power for humans will have similar predictive power for LLMs. I think that's a naive assumption, and that even if correlations exist the underlying causes are so divergent that the results are difficult to operationalize. So I'm not really sure what to do with studies like this until I find an ethical business model that allows me to make money by automating Mensa style test-taking en masse. Which I kind of hope will ever exist, to be honest.
2. MTurk is a hit mess (typo, but sic). If you want to do this type of study just recruit human participants in the old fashioned ways.
But given the goal of the authors, I don't think applying MTurk filters is "unfair". In fact, if anything, they're probably not doing enough.
I had no idea what was going on under the hood but used PGAnalyze for about 6 months after launching a new product. It was excellent at suggesting indexes and monitoring unused ones.
However, after a few months the ROI crept down until it wasn't worth it anymore (access patterns stabilized). I'm tempted to bring it back once and a while but the price tag keeps me from having it always on.
That makes sense, if the structure of your queries is largely static per commit of your codebase. You'd probably get more benefit out of a tool like this by running it as part of a CI pipeline, since that's the best time to test code changes including new queries. But then the challenge becomes simulating user activity and traffic levels during the CI pipeline. That's a solvable problem, but it's not as easy as just observing the real traffic in your prod network.
Exactly the same experience. And I wish it wasn’t. I also believe the pricing is wrong. I’d continue using it at $100/mo. But otherwise I’m just cancelling.
You can find our pricing on the website [0] - generally $100/month per instance (though note our smallest plan is $150/mo). pganalyze Index Advisor, which is the subject of the post is available with all plans.
Generally I (Founder and CEO) feel the pricing is fair for the value provided for production databases, and it allows us to run the business as an independent company without external investors, whilst continuing to invest in product improvements. That said, it may not be a good fit if you have a small production database, or only make database-related changes infrequently.
Seeing as you're here, the reason the price is too high for us is that "SCALE" plan is too much for us and the "PRODUCTION" plan is too little. I'd happily pay $200/month for two servers, but can't justify $400/month for two (we only have two right now).
Thanks, thats good feedback! And agreed, two servers is a bit of a gap in our pricing right now. If you're open to paying annual instead of monthly, we have a bit more flexibility since we can set them up as custom contracts (feel free to reach out)
My favorite thing about this report is on energy modelers (section 2.5). They show the (now famous) graph where large institutions (like the IEA) regularly predict that solar will start to have linear growth rather than exponential due to constraints in our energy system. This is despite every previous forecast that did this being wrong, with the number of new solar installations growing exponetially YoY instead.
What is interesting is that despite these constant misses, the IEA is constantly used as the definitive source for energy modeling across the world.
This along with pydantic [0] means that 2/3 of my favorite python open source projects are now commercially backed. I wonder how long FastAPI will last?
As an aside, what is the issue with versioning these days? Ruff and FastAPI both have massive user bases and a reliable codebase, but haven't released v1.0.0! Ruff hasn't even got as far as v0.1.0.
More seriously though, for these projects, the first version is version zero. If they make no major backward incompatible changes, why would they ever release a version 1?
This whole concept is so reckless in realms where the image content actually matters and people keep doing it anyways. You cannot CREATE information. You can infer it in certain situations, but if you infer the information and then analyze it you are setting yourself up to make mistakes by overextrapolating a bias/trend in your data to images where you have no idea if that inference is valid.
This was a big thing in the medical imaging community (where I did my stint as a CV researcher), folks were hallucinating microscope images and CT scans with no information theory justification as to why it worked.
Super resolution IS possible, but it must be done by synthesizing new pieces of information, not by inferring based on what other similar looking objects looked like. A cool technique by my former advisor does this with microscopes [1].
Deep learning has a place here, just not as a "lets create information" step, but as a way to learn how to synthesize additional information about images from more sources (i.e. more similar to how Google does Night Sight [2]).
Edit: if you want to see (an attempt) at using deep learning in this field you can checkout one of my papers [3].
This is all very sensible criticism but a bit generic.
Sometimes detail accuracy doesn't matter but the presence does.
Just about every image you ever view has had some manipulation applied. Sometimes that results in a "better" image.
Consider all astronomical images for human consumption, even smartphones adapt now to skin tone.
I'm playing hogwarts legacy, a recent AAA game which is very demanding, and where aesthetics are very important on a mediocre PC precisely because FSR from AMD (and if I had an Nvidea GPU DLSS and DLAA).
I should have been more nuanced I suppose. There is a time and place for these kinds of image "enhancements", they just don't belong in ESRI's scientific GIS platform. Folks don't view these images for pleasure (or at least very few do), they are typically used to analyze the satellite data or georeference other imagery.
Deep learning image enhancement is totally appropriate in your smartphone, as there the goal is not accuracy but perceived quality. Doing this to satellite imagery where the primary consumer cares about accuracy is what I call "reckless"
just wait for the 'real time enhancement' of drone cameras for critical security applications. Snark aside, it is very reasonable and competitive to want to do color correction, focus and shadow darkness on-the-fly; secondly raw sensing data is very large, but on-the-fly capture is bandwidth-sensitive, so very clever compression and band reduction is also desirable and indeed competitive.
What happens if there are multi-million dollar economic outcomes depending on the details of the remote sensing content, as in disaster response.
Agree, you can even see on their input/predicted/target examples that the created/invented data is off enough from ground truth to be in some cases unsuitable for photo interpretation.
Also back when I had to do these (I used Wave) having a notes section was very useful to include a few things (i.e. I used to include conversion rates). Would probably be pretty easy.