Post Opus 4.5, this pundit who does not write software has concluded that the case for coding agents has gotten weaker, not stronger, to the point where we might as well declare the debate settled in his favor. It takes him 8,000 words to do so.
Not before coding agents nor after coding agents has any PR taken me 5 hours to review. Is the delay here coordination/communication issues, the "Mythical Mammoth" stuff? I could buy that.
The article is referring to the total time including delays. It isn’t saying that PR review literally takes 5 hours of work. It’s saying you have to wait about half a day for someone else to review it.
Which is a thing that depend very much on team culture. In my team it is perhaps 15 min for smaller fixes to get signoff. There is a virtuous feedback loop here - smaller PRs give faster reviews, but also more frequent PRs, which give more frequent times to actually check if there is something new to review.
If I'm deep in coding flow the last thing I'm going to do is immediately jump on to someone else's PR. Half a day to a day sounds about right from when the PR is submitted to actually getting the green light
Similar in my team and I don't feel like there's much context switching. With around 8 engineers there's usually at least one person not in the middle of something who can spare a few minutes.
The PR won’t take 5 hours of work, but it could easily sit that long waiting for another engineer to willing to context switch from their own heads-down work.
Exactly. Even if I hammer the erstwhile reviewer with Teams/Slack messages to get it moved to the top of the queue and finished before the 5 hours are up, then all the other reviews get pushed down. It averages out, and the review market corrects.
Exaxtly. Can you get a lawyer on the phone now or do you wait ~ 5 hours. How about a doctor appt. Or a vet appt. Or a mechanic visit.
Needing full human attention on a co.plex task from a pro who can only look at your thing has a wait time. It is worse when there are only 2 or 3 such people in the world you can ask!
The article specified wall clock time. One day turnaround is pretty typical if its not urgent enough to demand immediate review, lots of people review incoming PRs as a morning activity.
I've had PRs that take me five hours to review. If your one PR is an entire feature that touches the database, the UI, and an API, and I have to do the QA on every part of it because as soon as I give the thumbs up it goes out the door to clients? Then its gonna take a while and I'm probably going to find a few critical issues and then the loop starts again
I use a PR notifier chrome extension, so I have a badge on the toolbar whenever a PR is waiting on me. I get to them in typically <2 minutes during work hours because I tab over to chrome whenever AI is thinking. Sometimes I even get to browse HN if not enough PRs are coming and not too many parallel work sessions.
One pattern I've seen is that a team with a decently complex codebase will have 2-3 senior people who have all of the necessary context and expertise to review PRs in that codebase. They also assign projects to other team members. All other team members submit PRs to them for review. Their review queue builds up easily and average review time tanks.
Not saying this is a good situation, but it's quite easy to run into it.
Life expectancy is not a useful health care system comparison because the primary factors that cause divergences between developed countries aren't based in the health system --- they're things like traffic accidents, homicides, drug overdoses, and suicide. Yes, CVD will appear in that list of factors, but it's noisy; despite having structurally the same health care system, states in New England will have Scandinavian CVD outcomes while southern states (some of whom actually do a better job than New England at making care available) have developing-nation CVD outcomes.
I buy that the locus of American overspending is in fees charged by providers (my understanding is that a further principle component of that spending is in end-of-life care).
The problem, though, with going after pharma costs, and pharma benefit managers is that pharma is a relatively small component of overall spending; it's less than 10%. That is to say, you could make all pharmaceuticals entirely free, and we'd get at best a 10% discount on what we pay. I don't think any of us would be satisfied with that!
This is data from the most recent (as of last year) CMS NHE:
The 9.2% figure is pharma's direct share of NHE, but drugs are a net-positive externality. Cheap statins can stave off cardiac surgeries, GLP-1 can stave off bariatric surgeries, etc. It's ridiculous to conclude we would only save 9.2% on costs--this is not zero-sum.
No comment on drug pricing and its incentives, the existence of America's prescription drug markets drives the new innovative drugs that the rest of the world picks up for cheap.
Mostly not marketing (still large), but the R&D costs and clinical trial costs. The latter are in hundreds of millions to billions range for the entire journey from a promising discovery to an FDA-approved medicine.
Every time Ive looked into it marketing is more than half of the costs of US pharma companies - and I would suspect even more as don't know if there has much work to unmask even more of that spending via channels that can occur in ways not obviously marked as marketing or at least are really not core to research and manufacturing.
e.g. is all the "discount coupon" pharmacy rigamarole considered marketing or administration.
This is not correct. Here's Pfizer's 2025 annual report [1]. Total expenses for the year were $55.1 billion. Advertising expenses were $2.7 billion of that, or just under 5%. R&D expenses were $12.1 billion, or just under 22%. They do have a lot of SG&A, but the large majority of that is not going to marketing.
Advertising is only a subset of marketing. From that doc, look at operating costs: SGA was ~$11B and R&D ~$12B - basically 50/50. Pfizer is very international, so is pretty difficult to break out US operating costs and what marketing vs R&D is for just the US. But one can also assume US marketing is higher than any other nation as direct-to-consumer advertising is primarily only allowed in the US.
No. Marketing is an issue but it's not the main driver.
Everybody else uses price controls to keep prices "reasonable"--the drug companies tolerate this so long as selling to the country exceeds their marginal cost of production. They count on the US market to recoup the $1B R&D costs.
Simply mandate that a drug company can't charge more in the US than they do in any other first world country. Major earthquake in drug costs.
The "discount card" bit is basically a reduction in revenue, it's neither marketing nor administration.
Right, but the idea that Americans specifically should pay higher prices is beyond propaganda. It's Stockholm Syndrome-level delusion. Big Pharm has thrived for generations on our research universities (for the time being anyway) and had a front row seat to expanding foreign markets under US-led globalization. In return, we get the world's most expensive healthcare system and the privilege of paying too much for meds because our leaders won't cut a deal. All they have to lose is the "hundreds of millions to billions range" in annual lobbying expenditures by Big Pharma.
In a sane world - or literally any other country - that $300-$500 million in annual lobbying would be the literal difference that makes medicine accessible for those who need it. Instead, it goes to expensive lunches.
Agreed. I mean, where did the COVID mRNA vaccine come from? Which company makes the GLP-1 inhibitors like Ozempic and Mounjaro? Are these American companies?
That's an interesting argument --- that massively increased access to pharmaceuticals would have knock-on impacts on other cost areas in the NHE.
I think if we dig into the numbers we're likely to find those effects, even if we maximize them, are marginal, unless we do other structural things to untangle the provider pricing system and do price transparency. Like: you could posit a material impact on CVD costs by making statins more widespread, and that should make a dent somewhere, but I don't know that CVD costs in non-Medicare-insured patients are really that big a line item, and non-Medicare is important here because people already Medicare-qualified generally have all the statins they want already. Meanwhile, providers are still ripping patients (and insurers) faces off for shoulder impingements, stents, and spinal fusions.
Wait, I hate employer-provided health insurance and think it's a terrible policy but what does that have to do with providers charging everyone --- including Medicare! --- way too much for services?
It’s a round about recognition of the agency problem in the medical industry.
If people chose and directly paid for there own medical bills and insurance then extra fees and extra diagnostics would be born directly by the person paying for it, who would have the freedom to make other choices, like picking insurance providers who were better at preventing it.
At least that’s an argument you can reasonably make. I’m not sure it would hold up in practice given how different medicine is from other markets.
The health insurance industry drives highly increased administrative costs - costs which the insurance companies are happy to foist off onto non insurance channels?
When one of my kids was 4, they had an unexplained seizure. Hospital workup, whole nine yards, never recurred; it was probably a medication reaction. Years later we were denied coverage from all the private insurers over it (more accurately: we were denied any coverage for that child).
Similarly, insurers would as a matter of course exclude from coverage any woman with one of several extremely common conditions, including endometriosis, PCOS, fibroids, and adenomyosis.
Prior to Obamacare, insurers were free to deny coverage wholesale for these conditions. It would have been fucked up to extend coverage but exclude any neurological conditions from my kid, but the actual outcome was worse: they were under the law entitled to withhold any coverage.
Before ACA, insurance had a more traditional "insurancey" role by excluding pre-existing conditions (aka managing moral hazard) in order to make money via premiums. In the "guaranteed issue" world post-ACA, insurance companies have pivoted instead to extracting as much money as they can from an increasingly vertically integrated ecosystem (PBMs etc)
Before ACA, insurance had a more traditional "dump you when you were in need" role that leveraged pre-existing conditions rules by, if you fell ill in a way which was likely to be sufficiently expensive to make this profitable, looking for any minimally tenable evidence of an undisclosed pre-existing condition (just to have something to cite as a pretext, it didn't need to be convincing), using it to justify cancelling your insurance, and avoiding any legal remedy by refunding your entire lifetime of premium payments.
Of course, whether or not you actually had a pre-existing condition when you had signed up for the original insurance, you now have one that prevents you from getting new insurance,
The problem here is that pre-ACA you didn't have insurance, either.
Yes, they would (maybe--some plans saw huge sticker shock because the original didn't actually pay much of anything) sell you "insurance". They would offer a plan for a few years, then close it, offer something new. The old plan would see patients getting sick, costs would rise. Premiums were based on costs *for that plan*. Soon the healthy jump ship for something else, now the old plan is only the sick and the premiums go into a death spiral.
Thus the reality was that any ongoing problem soon you were uninsured.
You have the two mixed up. Insurance companies - even for group insurance like through your company where they always had to accept everyone - required you to have “continuing coverage” and not have gaps or you had waiting periods.
The ACA also was written to enforce that through mandates and subsidies - a carrot and stick approach. The moral hazard was caused once there weren’t any mandates because of lawsuits by Republicans and the insurance companies still had to accept everyone.
We are talking past each other. When I said moral hazard I'm talking about adverse selection(1), my bad if that was unclear. And I was responding to some comments about denial for pre-existing conditions, which as you point out is irrelevant for continuous coverage group policies. Removing the mandate penalty without also adjusting the pre-existing condition protections did introduce adverse selection in the current regime. None of that disqualifies my argument about the current state of affairs though re the business model of modern insurance companies
If you were just looking to shout "dems good GOP bad" and find others who agree with you, that's cool too.
It's further the case, regarding Medicare expense ratios, that their admin costs are low relative to private insurance because the median private insurance customer incurs far lower medical costs, leaving admin as a higher fraction.
It can't be both: either insurers are incentivized to authorize as much care as possible so as to fit more money through the 20% opening, or they're incentivized to deny care to minimize their expenses. Which is it?
What do you mean elevators go up and down clearly someone only wants to go in one direction. If they are below 80% they want costs to increase, over 80% costs better decrease.
The ideal long term strategy is to drive everyone’s costs to go up slowly over time slightly faster than inflation. Adding administrative burden to medical institutions is a perfect way to achieve that, but clearly that never happens…
Pharmaceuticals are a small component of overall US health spending. Upcoding is endemic across the entire system; it's endemic across the whole system. Ironically, the complaint you'd be making with upcoding under Advantage is that Medicare should be denying coverage to people; Advantage upcoding involves altering risk scores to authorize more care.
reply