stevenbedrick
So, an analogous "Word-killer" today would presumably have to implement all of the docx format's weird quirks etc. On the one hand, the file format is standardized and open, so in principle that should be possible; on the other hand, it's a pretty gnarly file format, with a lot of nooks and crannies. Ironically, I remember hearing once that some of the weirder nooks and crannies of the docx format have their roots in... Word's WordPerfect interoperability features.
And as somebody who recently spent far more time than he expected to trying to reliably get data _out_ of a set of mildly-complicated docx files, I can report that the various fiddly details that the OP notes as being particularly important in the legal domain --- very specific details of paragraph formatting, complex table structures, etc. --- are a huge PITA to deal with when working with the docx format.
Second, the programs you’re referring to (“diversity” meaning only people with the right skin color etc.) used a very clear and well-defined model of “under-represented minority” that (in addition to members of this or that minority racial or ethnic group) included women, people from rural zip codes, people who were the first in their families to go to college, people who grew up on free/reduced lunch, and a bunch of other categories that have long been known to be under-represented in science. This definition has been around for decades and there’s been a ton of research about the importance and efficacy of these kinds of programs, some of which are training programs to try and diversify the incoming pipeline of scientists, and some of which were additional funding pools to try and address the very, very, very well-documented gap in early career funding awarded to URM scientists. Characterizing it as just being about skin color is simply inaccurate.
Some of the other “DEI” research that has been terminated consists of projects looking at health issues that are specific to certain populations; the NIH has officially said that health disparities research is still allowed but we are all scratching our heads trying to figure out how to write grants for it that don’t run afoul of the current “rules” which are frankly nonsensical and very “vibe” based so it’s hard to know what will and what won’t be allowed. Almost like they don’t want us to bother trying…
You are of course correct that it’s utterly bananas that in 225 we are wasting our time trying to guess which words shouldn’t be used in grant applications, and that our colleagues at the CDC and NIH are having to do the same for their papers.
Usually it's not all that much effort to glance over some other project's documentation to figure out how to integrate it, and as to creating working code from an idea or plan... isn't that a big part of what "programming" is all about? I'm confused by the idea that suddenly we need machines to do that for us: at a practical level, that is literally what we do. And at a conceptual level, the process of trying to reify an idea into an actual working program is usually very valuable for iterating on one's plans, and identifying problems with one's mental model of whatever you're trying to write a program about (c.f. Naur's notions about theory building).
As to why one should do this manually (as opposed to letting the magic surprise box take a stab at it for you), a few answers come to mind:
1. I'm professionally and personally accountable for the code I write and what it does, and so I want to make sure I actually understand what it's doing. I would hate to have to tell a colleague or customer "no, I don't know why it did $HORRIBLE_THING, and it's because I didn't actually write the program that I gave you, the AI did!"
2. At a practical level, #1 means that I need to be able to be confident that I know what's going on in my code and that I can fix it when it breaks. Fiddling with cmakes and npms is part of how I become confident that I understand what I'm building well enough to deal with the inevitable problems that will occur down the road.
3. Along similar lines, I need to be able to say that what I'm producing isn't violating somebody's IP, and to know where everything came from.
4. I'd rather spend my time making things work right the first time, than endlessly mess around trying to find the right incantation to explain to the magic box what I want it to do in sufficient detail. That seems like more work than just writing it myself.
Now, I will certainly agree that there is a role for LLMs in coding: fancier auto-complete and refactoring tools are great, and I have also found Zed's inline LLM assistant mode helpful for very limited things (basically as a souped-up find and replace feature, though I should note that I've also seen it introduce spectacular and complicated-to-fix errors). But those are all about making me more efficient at interacting with code I've already written, not doing the main body of the work for me.
So that's my $0.02!
From the article:
> State law does not require a woman to notify authorities when she miscarries, but Truman said that women who miscarry in West Virginia can protect themselves against potential criminal charges by reporting the miscarriage to local law enforcement. > > “Call your doctor. Call law enforcement, or 911, and just say, ‘I miscarried. I want you to know,'” advised Truman.
Now, in fairness, the state's association of prosecutors has clarified that this is not official policy and that the association as a whole does not support it. But it's definitely not out of the realm of possibility.
https://www.newsandsentinel.com/news/local-news/2025/06/pros...
So while I am as sick as anybody about hearing about him, on a thread about advances in cancer treatments, it would be awfully weird to ignore the elephant in the room. Just sayin’.
And this isn't for "DEI" reasons, it's literally because for decades there used to be drug trials that excluded women and as a result ended up releasing drugs that gave half the population weird side effects that didn't get caught during the trials, or just plain didn't work as well on one group or another in ways that were really hard to debug once the drug was on the market. That was legit bad science, and the medical research world has worked very hard over the last thirty years to do better. We are admittedly not there yet, but things are a lot better than they used to be.
For a really interesting take on the history of racial exclusion and bias in medicine, I recommend Uché Blackstock's recent book "Legacy: A Black Physician Reckons With Racism In Medicine" which gave a great overview.
Oh! And also everybody should read Abby Norman's "Ask Me About My Uterus," it gives a fabulous history of issues around women's health.
This was a large program that funded general research into antiviral drugs- again, in preparation for next time.
Going on the experience of the last couple of weeks, and what we have seen in terms of grant and program terminations, I would not be surprised if they did a big grep for the substring "COVID" and used that as their starting point for finding things to cut, without actually paying attention to what the grants were really studying.
Part of that might involve being able to show some kind of financial analysis about what was being cut, to justify it and to get buy-in from congress and other relevant stakeholders, and to do the cuts in a way that minimized their impacts, gave everybody who was going to be affected adequate time to be part of the process, and to plan for how to manage their side of the situation.
Needless to say, what we are seeing now is… none of those things.
A good example of what a saner process might look like would be the federal workforce reductions that followed a big analysis on government efficiency that Al Gore and his team led during the first Clinton presidency; look up the Federal Workforce Restructuring Act of 1994 to see how it all went down. They spent six months making a plan, then got it through congress to fund buyouts (it passed with major bipartisan support).
What is overhyped is the actual “savings” that they are producing with all of this.
That’s the thing that isn’t being reported well, IMHO: federal grants aren’t just gifts that the government gives universities, they are contracts for universities to perform services (conducting very specific research programs with well-specified deliverables, running very specific educational activities for carefully defined populations of learners, etc.).
1. In absolute terms, the amount of money we're talking about (e.g. NSF+NIH) is a drop in the bucket compared to $2T; it's been well-covered elsewhere, but scientific research funding simply isn't a significant contributor to the federal budget (and its deficit). It's a rounding error next to defense, social security, medicare, and medicaid.
2. Basic scientific research is an investment in our grandchildren's lives and health, and one that (as described in the article) historically has resulted in a very good rate of financial return.
3. When trying to cut costs, it is important not to be penny-wise and pound-foolish. These decisions are difficult, and if we want to have a hard look at what science gets funded, that's fine... but we need to do that in some kind of organized, serious, and systematic way, and that process absolutely has to involve the people doing the work at some level, as well as (ideally) other stakeholders (e.g. the people who might be affected by whatever the research would be accomplishing). A chainsaw is not the right tool for that job.
As to the second question, about priorities and agendas, this one's a bit trickier. First of all, let's be clear: we are not seeing a "steering" of research funding. What we are seeing is more of a "indiscriminate slashing" based on truly nonsensical and heavily politicized grounds: - https://www.npr.org/2025/02/13/nx-s1-5295043/sen-ted-cruzs-l... - https://www.science.org/content/article/nsf-reexamines-exist...
Besides funding of new research grants, the bigger impact is actually in education, training, and workforce development. Training grants are being badly affected, and because of the threats to indirect costs and the general uncertainty about future research funding levels, many universities are dramatically decreasing their number of PhD admissions this year, or eliminating them altogether. I also know of several programs aimed at scientific workforce development that are on the chopping block, and we are worried about our existing students losing their funding mid-way through their PhDs. Again, that's not "steering", unless you mean "steering" in the sense of "steering American science off a cliff". We are, in the most literal possible sense, eating our seed corn. I'm reminded of that reservoir in California that was recently ordered to dump a bunch of water, which means that the farmers in that part of the state won't have as much as they are going to need this summer. That's what is happening in science right now, as we speak.
Secondly, I don't quite know how to say this, but it's not really the executive's job to decide which grants get funded. That's up to the scientific and programmatic leadership at funding agencies, together with the existing and well-defined scientific peer review process. We did indeed just have an election, but our system of government is very much not a dictatorship. This is literally part of why Congress set up dedicated agencies to administer scientific and medical research; those projects need to persist beyond any single election and need to be insulated somewhat from day-to-day politics. Congress making those funding decisions was also an expression of democracy in action.
Also, this is why funding agencies are staffed by career civil servants with domain expertise, and why funding agencies have extremely well-documented and carefully designed peer review processes in place, and why scientists from all over the country give up lots of time to participate in those processes several times per year. Undermining all of this process and structure --- which is in place to try and mitigate biases and ensure as much objectivity as is possible in an inherently subjective world --- is one of the best ways I can imagine to ensure that government-funded research becomes more ideologically-driven, not less.
Anyway, that's my two cents!
And yes, many of the examples I listed are there for regulatory reasons, and that’s a good thing. We have laws around IRBs for good reasons, and it’s very important to have professional support in making sure we are doing things the right way in that regard. Data use agreements are important- when subjects share their personal data with me so I can study it, they do so with the understanding that it will be handled properly and part of how we do that is via data use agreements, and we need professionals to help with that because I certainly didn’t learn enough about contract law in grad school to do a good job with it on my own.
There is obviously a conversation to be had about whether a particular regulation is appropriate or whether there’s too much of this or that red tape, and I think every scientist would be able to tell stories of administrative annoyance. But it’s absurd to argue that the solution is to burn it all down indiscriminately, which is what we’re seeing.
These are all people who are at extreme risk of losing their jobs in the next weeks and months because of the chaos happening with NIH funding, and I can say with certainty that I as a scientist and an educator am far more effective because I have these professionals working with me. This is what our indirects cover and it is absolutely crucial.
To your point about being blind to what many around me are thinking- by definition, I wouldn't know if I was, right? So I'm not going to try and argue one way or another about that. I will say, however, that I have worked with many colleagues with whom I disagreed about many different things, some of which fall under the general umbrella of what one might call "identity politics", and as a general rule have been able to have open and civilized conversations with them. One thing I have learned is to not make any assumptions about what somebody does or doesn't think about a given topic, as basically every time I've done that I have been surprised.
No "DEI" boogeymen in there.
When I talk to people outside academia about “DEI”, it’s clear to me that whatever they think that term means, it has no relationship to anything I’ve ever seen in my career (involving faculty searches, recruitment of students and staff, education, involvement in clinical trial design and recruitment, etc.).
One of the reasons for this last bit is that many users of assistive technologies have very dynamic situations. For example, if they have a progressive condition like ALS, their mobility and other capabilities and needs will change over time; even if the user doesn't have a progressive condition, their living situation very likely will change over time, which may involve needing to re-work their setup. This might involve something as major as moving from one care facility to another, or having to adjust to the presence of a new pump or something that causes EM interference with your device.
Even for somebody with a very static medical condition and very stable living situation, their computing setup won't be static, so will need ongoing maintenance; or, alternatively, their computing setup will become static if they end up dependent on a particular piece of assistive technology, which can lead to a whole other set of issues (which is how you get people running ancient versions of Windows for years and years and years, because it's what their eye-tracker works on, and the vendor is long out of business and so never got it working on Windows 10).
The other piece to the puzzle is that, speaking as a gross generalization, users of assistive technologies tend not to be wealthy people (many speech and motor disabilities make it challenging to work). So, you've got an expensive product, often involving ongoing operational costs, with zero economy of scale, for a small and relatively poor customer base; this all adds up to what may fairly be described as "a very challenging business". Adding another layer of complexity, the economics of commercial assistive tech in the US have a lot to do with insurance Medicaid rules; the only reason there is any commercial activity at all in assistive technology and AAC is because under some circumstances Medicaid will pay for a device, within certain very rigid and complicated boundaries. So now your business also is going to have to juggle that problem in order to keep the lights on, which will add another layer of constraint to what is already a complicated situation.
On the other hand, when the right technology is paired with the right user and the right support, the impact on the quality of life for the user and their family can be truly enormous. Which is what keeps people working in the space...
"Wisdom" is probably the hardest part of the framework to define clearly, but across formulations it's about being able to integrate different kinds of knowledge from difference sources, and then using that to be able to know when and why to take a certain action, or to know what the right thing to do might be. Obviously this is something that is frequently informed by knowledge derived from science and scientific reasoning, but usually also relies on some of the ways of thinking and understanding that the humanities can help provide. So that's how I'd understand what Devereaux was saying (allowing for some rhetorical flourish on his part)- it's a "both-and" kind of situation rather than an "either-or"; in other words, science and humanities on their own are both necessary and insufficient.
And of course for a humorous take on the matter, SMBC is always a reliable source: https://www.smbc-comics.com/comic/2009-10-18
Zooming out beyond academia, Bret Devereaux, a historian whose writing appears with some regularity on HN, wrote a superb essay a few years ago on the practical case for the humanities and puts it better than I ever could, so I will link to his post: https://acoup.blog/2020/07/03/collections-the-practical-case...