For adapting kids toys to this purpose, there's also K'nex.
ChatGPT has subverted everyone’s predictions on automation. Just a few years ago, it seemed most likely that the manual, boring, and rote jobs would be automated—but in the presence of GPT and the other newest gargantuan deep learning models like DALL-E, it seems more likely that writers, artists, and programmers are the most vulnerable to displacement. Everyone’s freaking out about it,
including me, except mine is more of a cynical freak out: I don’t want to live in a world where AI content is ubiquitous and human content is sparse and poorly incentivized—if only because the professions of a writer, artist, programmer etc. are some of the most fulfilling vocations out there. If the technological trend continues, we’re facing a future world where intellectual work no longer exists. This is the worst imaginable end-stage capitalism dystopia, in which the only ways to make money are the grueling physical jobs like nursing and commercial kitchens (if you work in a field like that, you have my deepest respect).I don’t think a language model can replace a programmer—it can only convincingly fool a hiring manager that it can. (Or I don’t know, maybe it will take like two more years of progress before the hiring managers are fooled. It managed to fool this guy.) And the same is true with writing and art—ChatGPT can’t actually replace a good human writer (yet), but it can certainly convince someone that it can do the job well enough for less money. It can certainly get literary magazines shut down by filling up its submission pipelines with its polite sludge. These generative models create a double-whammy of badness: the programmer will be out of a job and the company will find their infrastructure crumbling for every seasoned programmer they let go. Writers and artists won’t be able to make livings from their work and the content they’re not producing anymore will become horribly banal—where will we go to satisfy our curiosity then?
Whether you think ChatGPT is wonderful
or terrible, I hope we can agree on this: people should have the right to control whether the things they create are used to train the massive AIs of the massive for-profit tech corporations. You may think that OpenAI doesn’t need permission to use creations that are publicly available on the internet, but hopefully you agree that a person should be able to disallow OpenAI from using the things he or she creates, even if he or she wants to share those things with the world. Anyway maybe OpenAI should need permission to use creations that are publicly available on the internet.I think this is the direction that discussion and policy needs to move now that generative models are becoming ubiquitous. It is already at best questionable whether OpenAI et al. should be allowed to use online content without permission from the creator. And it is already at best questionable whether ChatGPT and the like represent a net good for society—even putting aside the potential existential risk to humanity. There needs to be some sort of regulation on this newish industry which feeds off of all of us, and it needs to be enforced. So the question has to be what sort of regulation is fair & feasible?
There’s been talk of compensating people for their contributions to the training data, which is a nice, utopian, idea, but it’s not feasible to implement. How do you track down everyone that contributed? How do you determine the value of their contribution? And do you really think that the compensation would be anything more than pennies?
On the more pessimistic end of things, there’s the call to halt AI research. Keep in mind that this isn’t a call to make it illegal to use neural networks, it’s only a call to stop creating neural networks at the scale that’s at the limits of our capabilities. It’s basically a call to hold off on GPT-5. I think this would basically be great. I don’t think we’d really be giving anything up. There are genuinely good applications of “AI” like deep learning for protein folding (that’s biology research using AI, not AI research; AI research is about the bleeding edge of AI itself), but I don’t think anyone is calling for a pause on that. I don’t think ChatGPT and the like are profoundly useful or good for society. I think they’re mostly harmless, but then what about the next generation of GPT & co.? What sort of consequences will we be dealing with in five years, or even next year, if these developments continue unchecked?
Alas, it’s unlikely there will be a halt on AI research, even if half of the world population signed that letter. OpenAI, Google, and Microsoft aren’t just going to say, “Okay, you’re right everyone, let’s shut it down,” and I’m not optimistic that the government would be able to enforce a full stop, even if we were facing a more immediate, obvious, and severe threat.
In the middle, there’s this idea, which I haven’t seen discussed much: we should all simply refuse to be included in training data, or at least we should all have the right to refuse to be included in training data.
If you’re an individual, and you’re worried about all this, all you have to do is either explicitly disallow the things you produce from being included in AI training datasets, set some platform setting that does this for you, or get off the platform altogether if it does not give you this freedom.
If you’re a governing body, all you have to do is go to OpenAI, Google, and Microsoft (importantly, only these and maybe a couple other organizations are really the only ones capable of the sort of AI training we’re talking about) and require that they obtain the explicit permission from the creator of every observation in their dataset. (I know nothing about corporate policy. Maybe this could never be enforced and is, in general, terribly naive. But it has to be easier to enforce than an absolute shutdown of AI research or paying people for their contributions to datasets.)The way these “innovations” in AI like ChatGPT work is basically that they amass absolutely disgusting quantities of data. How disgusting, you ask? Well, OpenAI doesn’t tell us how big their datasets are, or where the content comes from anymore. That’s how disgusting.
The actual innovation was in 2017, with what is turning out to be one of the most important papers of all time, titled “Attention Is All You Need.” I remember reading this paper over and over in 2018-2019 while I was implementing and training these models at work. Basically, the authors discovered that you could get a lot more bang for your buck out of neural networks trained on text data by constructing them entirely out of “attention mechanisms” (which the specifics of are relatively interesting if you work in machine learning but not interesting at all otherwise). This led to the development of the “transformer architecture,” which pretty much revolutionized machine learning for text (GPT stands for Generative Pre-trained Transformer—the first version was created in 2018. Before that, OpenAI was still working on reinforcement learning [note the date on that link—2017!], and their arena of choice was one of my all-time favorite video games, DotA 2).
(As an interesting aside: the paper “Attention Is All You Need” was so popular and important and its title was so cute and fun that researchers copied it to an irritating degree, even in fields outside machine learning: see “Diversity Is All You Need,” “A Lip Sync Expert Is All You Need”, “Empathy Is All You Need,” and 29,000 other results on Google Scholar.)
Since 2017, machine learning researchers have pretty much just been throwing more and more data at these transformer models and getting better and better results. It just turns out that the performance of this model architecture scales very well and very far out with the size of the dataset (notably, Sam Altman has said that he thinks we’re approaching the limits of this relationship—I’ll let you decide what to make of that).
Understanding the history here is helpful in two ways. Firstly, it gets you past the overly simple “All it’s doing is predicting the next word!” way of comprehending these models. While this isn’t overtly wrong, it would also be accurate to say “It’s a system, the complexity of which rivals that of a mammalian brain, that is somehow encoding a rich representation of the entire english language, and responds to written language with novel, relevant, and apparently intelligent language based on an extremely complex network of mathematical relationships which resemble, to some degree, the way humans process written language.”
But secondly, it exposes the fact that GPT-4 is more impressive that GPT-3 mostly just because the training dataset is bigger. It’s really not a function of clever programmers inventing the smartest AI. There’s some of that, yes, but the project is really more about getting a bigger dataset. Acquiring and maintaining training data is where most of the cost and effort for creating these models come from. If we introduce even a little bit of resistance on the capabilities of an organization to perform this data-hoarding, it will become a lot harder to produce these sort of models, because they’re operating at the limits of what is possible in terms of big data—it might even be impossible to make any more progress. You may think that is a good thing or a bad thing, but it’s hard not to feel like OpenAI is being shady by not disclosing any details about their training data, and maybe regulators need to step in and do something about that.
In fact, it’s already happening. Take a look at this WSJ coverage of recent legislation from the EU, and discussion on Hacker News. This is already the way things are headed, and that’s good, even if you think ChatGPT is good for society. What I’m advocating for is basically just an extension of privacy (which we’ve already basically agreed as a society is important, and which we’re already building policy and infrastructure for): AI privacy.
It seems like people want this. DeviantArt prohibited the use of content for AI datasets in response to user feedback. And look at all the downvotes on this Stack Overflow post, where the platform announces that content will be used in AI training.
Is putting a red cancel emoji in your Instagram bio going to stop OpenAI from downloading your art and using it in training data anyway? No. I mean, the Instagram terms of service could just say that you give up all rights by using the platform, and then they’ve got full legal freedom to sell your content to OpenAI.
Are people going to unite in rebellion against these platforms if they don’t refuse to share their data with OpenAI? No, probably not.
For instance, I don’t think people are going to stop using Stack Overflow. That doesn’t mean that their decision isn’t a bummer. That doesn’t mean we shouldn’t push Stack Overflow to reverse their position like DeviantArt did. That doesn’t mean someone couldn’t create a very similar programming Q&A site that doesn’t allow AI consumption, and see if people wouldn’t prefer to use that one. The platforms need the users—if everyone starts leaving because they don’t want AI to replace them, then they’ll stop sharing the content with AI—but the users also need the platforms, and platforms are hard to build, especially when you consider the requirement of strong network effects (good luck getting your Stack Overflow alternative off the ground: step 1 is getting everyone to leave Stack Overflow), so the platforms have at least half of the power in this dynamic. But this is the sort of way in which regulatory pressure actually does work: shifting the power balance toward the users, toward the creators. We probably can’t create regulations that directly prevent superintelligent or somehow malignant AIs from being created, but we can create regulations that pressure platforms to behave a certain way, and therefore improve the world by improving the way that these technologies are allowed to be developed.
As a developer, I often find myself knee-deep in a new technology – perhaps investigating a library, or learning a language. I’m trying to frame new concepts in my head, applying my own data and architecture on the fly to the generic explanations in the docs. It’s hard! Which is why it’s jolting to read something like:
- [This library] makes it painless to [do difficult thing].
- [Complicated thing] made simple and easy.
- All you have to do is just [difficult thing].
If someone’s been driven to Google something you’ve written, they’re stuck. Being stuck is, to one degree or another, upsetting and annoying. So try not to make them feel worse by telling them how straightforward they should be finding it. It gets in the way of them learning what you want them to learn.
When I’m writing technical docs, annoyingly, I find myself putting “just” in all the time, despite having strongly-held views about words such as this. I think it’s because I’m so keen to share my enlightenment. It’s exciting that the code I’ve written might be able to help, and naturally I’m keen to communicate that. A bit of light editing lets you keep the enthusiasm for the code, lose the bits that could be interpreted as condescending, and improve the clarity, to boot.
If you find smartphone notifications annoying enough already thanks to their skill at exploiting the full range of distraction options available, whether dropping a banner from above or sprinkling pox-like red balls over your homescreen icons so as to lodge like grit in the eye, you should prepare yourself for even less subtle demands bubbling into your eye-line in the future if novel research into flat panel haptics ends up being commercialized by mobile device makers.
Think notifications that create a physical bulge in the screen of your smartphone — making the update icon stick out or even pulse lightly like the proverbial sore thumb until you press with your own digit to remove the unsightly wrinkle.
On the less dystopian side, touchscreens with the ability to be dynamically tactile could have accessibility benefits by enabling form and texture to co-exist with the utility of flat panel computing — for instance by providing people with visual impairment with physical signals to help identify key on-screen content (paired with the necessary software to power such a use-case in existing apps and interfaces of course).
The ever inventive Future Interfaces Group at Carnegie Mellon University is behind the research into what they describe as “embedded electroosmotic pumps for scalable shape displays”. The main break-through they’re claiming here is squeezing the hydraulics-based haptics into a thin enough panel that it can be sequestered behind an OLED screen — such as those found on modern smartphones.
Their work is detailed in this research paper (PDF) — and demoed in the below video:
While bulging notifications might not be the average smartphone users’ idea of futuristic mobile computing heaven, the researchers suggest the prototype tech could allow for dynamic interfaces on other types of devices so that buttons and signals appear at the point of necessity — say power, play and track progress on a music player — rather than having to fit in lots of physical knobs and dials.
They also trail the idea of the flat panel haptics tech enabling the return of keyboard physicality to touchscreen smartphones.
Long time mobile industry watchers may recall that BlackBerry-maker RIM, a company which dominated the mobile arena in the pre-iPhone touchscreen era with its designed-for-email physical keyboard handsets, actually tried something like this all the way back in 2008.
The ill fated BlackBerry Storm, as the ‘turducken’ handset was named, combined a touchscreen with embedded physical haptics — the screen literally clicked as you pressed — in a bit to recreate the sensation of pressing real keys on a physical-Qwerty-free touchscreen handset.
The problem was, er, the experience basically sucked. It was neither fish nor foul, as the saying goes. So whether lots of mobile makers will be rushing to embed electroosmotic pumps into their handsets just to have another bite at keyboard physicality in the era of touchscreen computing seems debatable.
Although tablets seem a much more interesting use-case. (And, beyond that, the general idea of squeezing more attention-grabbing bells and whistles into roughly the same physical space will surely have takers.)
Add to that, RIM’s attempt to implement a touchscreen keyboard with physicality some fifteen years ago was clearly lacking the fine-grained tactility needed for the tech to perform usefully in a typing context, since the company apparently just stuck a single button under the screen’s backplate.
Whereas the researchers point out their electroosmotic pumps can be as small as 2mm in diameter (and up to 10mm), with each pump being individually controllable (akin to pixels) and supporting fast update rates. This suggests that a flexible touchscreen combined with an array of their miniaturized hydraulics could be a lot more dynamic and versatile (and thereby potentially useful) than was possible with the sorts of mechanical mechanisms available for pairing back in the day.
So there is still a chance that RIM’s BlackBerry Storm was simply ahead of its time.
Bob Widlar (1937-1991) is without a doubt one of the most famous hardware engineers of all time. In fact, it would not be an exaggeration to say that he is the person who single-handedly started the whole Analog IC Industry. Sure, it’s Robert Noyce and Jack Kilby who invented the concept of Integrated Circuits, but it’s Widlar’s genius and pragmatism that brought it to life. Though he was not first to realize the limitations of planar process and designing ICs like discrete circuits, he was the first one to provide an actual solution – µA702, the first linear IC Operational Amplifier. Combining his engineering genius, understanding of economic aspects of circuit design and awareness of medium and process limitations, he and Dave Talbert ruled the world of Analog ICs throughout the 60s and 70s. For a significant period of time, they were responsible more than 80 percent of all linear circuits made and sold in the entire world.
The list of his designs includes gems such as µA709, improvement over original µA702 and a Fairchild’s flagship product for years, µA723 — first integrated voltage regulator and LM10 — the first ultra-low-voltage opamp, which is still in production today. Students usually learn about Widlar via the textbook-classic: Widlar Current Source, a key piece in many of his designs, and the Bandgap Voltage Reference – both of which provide an infinite supply of mind-boggling exam problems. If there is one theme that’s common across all of Widlar’s designs, it’s that he has never designed an obvious circuit in his life. Every Widlar design comes with a twist, a unique idea and very often, a prank. A classic example of this is the story of LM109, the industry’s first three-terminal adjustable voltage regulator. In 1969, Widlar wrote a paper in which he argued against feasibility of monolithic voltage regulators due to temperature swings and packaging limitations. Since he was already an engineering legend by that time, the industry took it seriously and people gave up trying to pursue such devices. Then in 1970, he presented a circuit — LM109 — which used his bandgap voltage reference to achieve exactly such “impossible” functionality. It is most likely that he submitted both works within days from one another.
In addition to being a brilliant designer, Widlar was a personification of an age to come in Sillicon Valley, combining counter-cultural and in-your-face attitude with entrepreneurial passion and desire to build products that people love. He worked directly with customers and wrote his own app notes and data sheets. In fact, Widlar’s µA702 laid out the blueprint for how all analog IC data sheets are to be written in the future. His principle was “designing for minimum phone calls” and “if you make a million ICs; you get half a million phone calls if they don’t work right”. He was both destroyer of the worlds and creator of new markets; he came into Fairchild claiming that “what they do in analog is BS”, but left the company as a dominant player in linear IC for years to come, mostly on the wings of his designs. He then moved to Molectro (owned by National) but quickly ended up turning the parent company upside down and making into an Analog powerhouse. At the age of 33 he cashed out and retired in Mexico. But his hands couldn’t stay idle for too long. He soon came back as a contractor for National and, in 1980, ended up founding Linear Technology with Robert Swanson and Bob Dobkin,
Still, he always remained a troublemaker, free thinker, and an HR nightmare… closer in spirit to someone like Hemingway than a fellow “professional” engineer. Such attitude was contagious and it inspired a whole new wave of “prankster” analog geniuses like Bob Pease and Jim Williams. Widlar’s pranks are too many to count and it’s really hard to pick one that captures the spirit of the times the best. Maybe it’s when Widlar brought sheep to the front of National as a reaction to the firm not mowing the lawns due to cost-cutting (he really just needed an excuse to annoy the upper management). Or when he cherry-bombed the intercom speaker, again, just to upset one of National’s Vice Presidents. Some of the pranks were actual hardware, like a “hassler” circuit he built to detect audio, convert it to a very high audio frequency and play back the converted sound. The net effect of such a design was that the louder someone talked in the office, the more annoying the “ringing” effect caused by the feedback was. As a person would stop shouting to hear what’s causing the ringing, the effect would disappear as well. This way, he managed to get everyone in the office into speaking quietly, Pavlov-style.
Widlar passed away in 1991 but his legacy lives on. He truly was the original hardware hacker and more than just an engineer – he was an Artist. It’s because of guys like him that Analog still has that special feel and is more about “invention” than just following the straightforward path between A and B. And that is why Analog guys still greet everyone else with a “Widlar Salute”.
Now, when I have finished my inspection, and I am still mad as hell because I have wasted a lot of time being fooled by a bad component – what do I do? I usually WIDLARIZE it, and it makes me feel a lot better. How do you WIDLARIZE something? You take it over to the anvil part of the vice, and you beat on it with a hammer, until it is all crunched down to tiny little pieces, so small that you don’t even have to sweep it off the floor. It makes you feel better. And you know that that component will never vex you again. That’s not a joke, because sometimes if you have a bad pot or a bad capacitor, and you just set it aside, a few months later you find it slipped back into your new circuit and is wasting your time again. When you WIDLARIZE something, that is not going to happen. And the late Bob Widlar is the guy who showed me how to do it.
Bob Pease – Troubleshooting Analog Circuits
[1] Bo Lojek – History of Semiconductor Engineering, Springer, 2007
[2] Bob Pease – Troubleshooting Analog Circuits, 1987
[3] http://en.wikipedia.org/wiki/Bob_Widlar
[4] http://readingjimwilliams.blogspot.com/2012/04/my-favorite-widlar-story.html
[5] http://analogfootsteps.blogspot.com/search/label/Bob%20Widlar
[6] http://electronicdesign.com/analog/what-s-all-widlar-stuff-anyhow
[7] http://silicongenesis.stanford.edu/transcripts/dobkinwilliams.htm
[8] http://edn.com/electronics-blogs/anablog/4311277/Bob-Widlar-cherry-bombs-the-intercom-speaker-item-2
Nix package manager on Android, in a single-click installable package. This is not full NixOS running inside Android, but you get easy access to nixpkgs' vast collection of (precompiled!) software and the best package manager under the sun. It's prototype-grade quality as of now, but hey, it works!
It does not require root, user namespaces support or disabling SELinux, but it relies on proot and other hacks instead. It uses a fork of Termux-the-terminal-emulator app, but has no relation to Termux-the-distro. Please do not pester Termux folks about Nix-on-Droid.
This repository contains:
nix-on-droid executable.It is only tested with aarch64 (64-bit ARM devices). It also used to compile for i686 devices, but the developers don't own any and nobody has reported whether it actually worked or not, so it's no longer built unless a user shows up. Sorry, it would not work on 32-bit ARM devices and it's not an easy feat to pull off.
Install it from F-Droid, launch the app, press OK, expect many hundreds megabytes of downloads to happen.
The Nix-on-Droid system can be managed through a custom config file in ~/.config/nixpkgs/nix-on-droid.nix as generated on first build, for example:
{ pkgs, ... }:
{
environment.packages = [ pkgs.vim ];
system.stateVersion = "22.11";
}An alternative location is ~/.config/nixpkgs/config.nix with the key nix-on-droid, for example:
{
nix-on-droid =
{ pkgs, ... }:
{
environment.packages = [ pkgs.vim ];
system.stateVersion = "22.11";
};
}See https://t184256.github.io/nix-on-droid/ for list of all available options.
To enable home-manager you simply need to follow the instructions already provided in the example nix-on-droid.nix:
home-manager channel:
nix-channel --add https://github.com/rycee/home-manager/archive/release-22.11.tar.gz home-manager nix-channel --update
home-manager:
{ pkgs, ... }:
{
# Read Nix-on-Droid changelog before changing this value
system.stateVersion = "22.11";
# insert Nix-on-Droid config
home-manager.config =
{ pkgs, ... }:
{
# Read home-manager changelog before changing this value
home.stateVersion = "22.11";
# insert home-manager config
};
# or if you have a separate home.nix already present:
home-manager.config = ./home.nix;
}nix-on-droid executableThis executable is responsible for activating new configurations: Use nix-on-droid switch to activate the current configuration and nix-on-droid rollback to rollback to the latest build.
For more information, please run nix-on-droid help.
The terminal emulator part is probably not interesting for you, just download and use a prebuilt one. If you really want to rebuild it, you can just use Android Studio for that.
The zipball generation is probably what you are after. Get an x86_64 computer with flake-enabled Nix.
tl;dr: Use the deploy app like the following which executes all steps mentioned below:
nix run ".#deploy" -- <public_url> <rsync_target> # or run the following for explanation of this script nix run ".#deploy"
Run
nix build ".#bootstrapZip" --impurePut the zip file from result on some HTTP server and specify the parent directory URL during the installation. To re-trigger the installation, you can use 'clear data' on the Android app (after backing stuff up, obviously).
If you want to change the Nix-on-Droid channel to your custom one, you can do that either with nix-channel after the installation, or by setting the environment variable NIX_ON_DROID_CHANNEL_URL. Other environment variables are NIXPKGS_CHANNEL_URL an NIX_ON_DROID_FLAKE_URL.
Note: The proot binary is not built on the android device (NDK is required for building it, and it's not available on mobile platforms). The way we work around it is to push proot derivation to cachix. The current workaround is to hardcode the path to the wanted proot nix store path in modules/environment/login/default.nix. During evaluation time on the android device this store path will be downloaded from the binary cache (https://nix-on-droid.cachix.org/). This in return means the proot derivation has to be present there or in any other binary cache configured in the nix.conf on the device.
Obviously it's an annoyance if one wants to fork this repo and test something. To minimize the hassle with this scenario, proot derivation is also bundled with the bootstrap zipball. This way you only need your own binary cache if you are planning to maintain a long-term fork that users can update from. In case you only care about updates through wiping the data, or are forking to submit a one-off pull request, you shouldn't need a binary cache for that.
Note: Nix flake support is still experimental at the moment and subject to change.
A minimal example could look like the following:
{
description = "Minimal example of Nix-on-Droid system config.";
inputs = {
nixpkgs.url = "github:NixOS/nixpkgs/nixos-22.11";
nix-on-droid = {
url = "github:t184256/nix-on-droid/release-22.11";
inputs.nixpkgs.follows = "nixpkgs";
};
};
outputs = { self, nixpkgs, nix-on-droid }: {
nixOnDroidConfigurations.default = nix-on-droid.lib.nixOnDroidConfiguration {
modules = [ ./nix-on-droid.nix ];
};
};
}For more examples and nix flake templates, see templates directory or explore with:
nix flake init --template github:t184256/nix-on-droid#advanced
nix-on-droidUse nix-on-droid switch --flake path/to/flake#device to build and activate your configuration (path/to/flake#device will expand to .#nixOnDroidConfigurations.device). If you run nix-on-droid switch --flake path/to/flake, the default configuration will be used.
Note: Currently, Nix-on-Droid can not be built with an pure flake build because of hardcoded store paths for proot. Therefore, every evaluation of a flake configuration will be executed with --impure flag. (This behaviour will be dropped as soon as the default setup does not require it anymore.)
fakedroidFor easier debugging and testing, there is a so-called fakedroid environment which emulates the on-device environment including proot as good as possible.
You can enter this emulated environment with
nix run --impure ".#fakedroid"This will install Nix-on-Droid with initial channel setup and save the state of this session in the .fakedroid directory in the project directory. To enter this environment with initial flake setup, set USE_FLAKE=1 as environment variable. Channel and flake setups can co-exist in the .fakedroid state directory. To rebuild/re-test the bootstrap process, remove the .fakedroid directory.
To only run a command in fakedroid and not enter an interactive shell, just call it with the command as arguments:
nix run --impure ".#fakedroid" -- ls -lIn the ./tests/on-device directory, there is small set of bats tests that can be executed on the android device or run in the fakedroid environment.
To run the tests (on device or in fakedroid), run
nix-on-droid on-device-test
Note: This currently requires a channel setup and should only be executed on clean installations.
OK, real brief.
Developer's device:
proot for the target platform is cross-compiled against bionic, (to fake file paths like /nix/store; think 'userspace chroot')nix is taken from the original release tarballnix database is initializednix and the Nix-on-Droid module systemUser's device:
You can refer to a NixCon 2019 presentation talk for a more extensive overview of the subject.
Licensed under MIT License, see LICENSE. Copyright (c) 2019-2021 Alexander Sosedkin and other contributors, see AUTHORS.
Two rewrites ago it was based off the official Nix install script (https://nixos.org/nix/install), presumably written by Eelco Dolstra.
Is deployed and used with a fork of Termux-the-terminal-emulator app, but has no relation to Termux-the-distro.
Previous project that did use Termux-the-distro: https://github.com/t184256/nix-in-termux
Share your rental experience with confidence.
Join fellow Tenants by creating an informed community.
Empower others to make decisions about housing.
The artificial intelligence (AI) chatbot, ChatGPT, is once again available to users in Italy after its owners addressed data privacy concerns, an Italian regulator said on Friday.
Italy blocked the site at the end of March after raising concerns about how ChatGPT processes and saves user data.
Garante, Italy's data protection authority, said ChatGPT has been reinstated "with enhanced transparency and rights for European users."
The move came after OpenAI, the US-based and Microsoft-backed company that developed the chatbot, implemented changes to comply with several data privacy conditions.
The changes include increased transparency on OpenAI's website for how the chatbot processes user data.
The platform now also grants users certain opt-out rights, including being able to toggle off the option for conversations to be used for training ChatGPT's algorithms.
The site now also has checks to protect children under 13, with age verification in place for those accessing the site from Italy.
"ChatGPT is available again for our users in Italy. We are delighted to welcome them back and remain committed to protecting their personal data," an OpenAI spokesperson said.
When viewing the platform from Germany, the new data protection options were also available.
The site also now has a notice making users aware that ChatGPT could produce inaccurate information about "people, places or facts."
Italy became the first Western country to take action against ChatGPT at the end of March.
The country's data protection watchdog said its developers did not have a legal basis to justify the storage and collection of users' personal data in order to train the site's algorithms.
Authorities also criticized that inaccurate information produced by the platform was not being handled properly and that children were especially vulnerable to "absolutely unsuitable answers."
ChatGPT has taken the world by storm since it was launched in 2022.
Based on questions or input from users, the chatbot can generate essays, poems, songs, computer code and also news articles.
While its release has been heralded as a milestone in technological advancement, it has also sparked a debate about the possibilities of artificial intelligence and the consequences that could arise.
rs/sri (AFP, dpa)
Welcome to Clubhouse Eats and Drinks, where we’ll introduce you to our favorite food and drink at some of the best golf courses, golf resorts and golfy locations around. Cheers!
For Southerners, pimento cheese has long been a ubiquitous presence at backyard barbecues and cocktail parties. For the rest of us, it’s a somewhat alien form of sustenance. What exactly is a pimento cheese sandwich anyway? Like most things, opinions vary on best practices when it comes to its preparation, but just about every iteration includes the following: a mix of cheddar cheese, mayonnaise and pimentos (mild peppers with a sweet flavor, in case you were wondering) piled generously between two slices of white bread.
A quick scan at the ingredient list on the wrapper of Augusta National Golf Club’s celebrated version reveals the additional inclusion of Monterey Jack cheese, cream cheese, onions, salt, pepper and cayenne pepper. Home chefs might opt to mix it up further by adding Worcestershire sauce or chile paste.
This week we’re going to take a detour into understanding the structure of academia, in particular the different kinds of ‘professors’ and their academic ranks in the American system, with a particular focus on ‘non-tenure track’ faculty (which is to say, as we’ll see, ‘most teaching faculty.’) This is intended as the first in a series of posts mixed into the normal diet over the next few months looking at the structure of the modern American university from the inside. The fact is that while quite a lot of people go to college, few students acquire much of any sense of how their college or university is structured, and so there is a tendency for a lot of folks to believe they know how academia works who don’t, in the same way that most people who eat at fast food restaurants cannot, in fact, operate their kitchens.
My own experience of course has been as a student, then a graduate student (worker), then as adjunct faculty at three different Big State Universities. Less so in this post, but more so in later posts I’ll also be drawing on the experiences of my better half a bit, as she’s been an administrative staff member for several academic departments and one research program across two Big State Universities and so has a lot of visibility into the bureaucratic structures involved. As you might guess with that background, I am going to be particularly focused on Big State Universities, but I actually think that is good – compared to the Ivies or Small Liberal Arts Colleges, Big State Universities make up the largest single chunk of 4-year-degree institutions and indeed grant a simple majority of 4-year degrees, so the Big State University is by raw dint of numbers both the median and modal higher education experience for folks who achieve a four year degree.
We are in particular going to focus on non-tenure track (NTT) faculty for two reasons. First, because while NTT make up the simple majority of student-facing teaching faculty, universities go to considerable length to obscure this fact leaving many students incorrectly assuming their professors are largely tenure-track when at many institutions they may not be. And second because I’m a NTT faculty member (who, like most NTT, would like to be on the tenure track for reasons which will become obvious below) and I wanted to explain all of this in one permanent place in part so I can point back to it, in particular because while NTT faculty members are the most common they are also the least understood by the public. But we’ll still talk a little bit about the tenured ranks too.
And as always, if you want to support my public writing on historical topics, you can help both by sharing what I write (for I rely on word of mouth for my audience) and by supporting me on Patreon. This blog is reader supported; for reasons that will become clear when we come to talk about adjuncts, I receive no research or writing support from the universities I teach for and so this project serves to fund both my public writing (here and elsewhere) and my research work as well. If you want updates whenever a new post appears, you can click below for email updates or follow me on twitter (@BretDevereaux) for updates as to new posts as well as my occasional ancient history, foreign policy or military history musings, assuming there is still a Twitter by the time this post goes live.
We need to start by outlining the structure of the university and all of its employees. Universities are very big. Even many small liberal arts colleges will have several hundred (if not many hundreds) of employees and large state universities have thousands; UNC-Chapel Hill has 19,743 undergraduates and 12,961 total staff members, for instance. I should note that while there are many small liberal arts colleges (SLACs) in the USA, the enormous size of large, public R1s means that collectively they make up more than half of the US university system by both faculty and students, so this is a case in which the big schools have become typical because they are so big to swamp everything else. That said, smaller institutions matter and what I am going to say here should apply broadly; I will note where conditions differ for different kinds of institutions.
So let’s start dividing all of those employees down so we know what we’re dealing with. We can start by splitting the university into faculty and staff (with student-workers as a third group we’ll not discuss this week); faculty teach and do research whereas staff are all of the supporting administrators and workers that make the university function. We’re not going to talk much about staff, but briefly we can divide them quickly into four big groups: leadership (chancellors, deans, and assistant deans of various kinds; of old these used to be professors pulled into leadership temporarily but these days these are professional managers), department staff (who work within academic departments handling the scheduling, paperwork and other essential support services), university staff (who staff the university-wide bureaucracies like the registrar or bursar) and finally what I’ll call – somewhat imprecisely – facilities staff (a wide category covering all of the folks who do a lot of the physical work that keeps a university running; repair, grounds-keeping, janitorial tasks, running dining areas, etc. etc.). All of these people are important, but this week’s post isn’t about them; I break them up here so that when I do mention them, you understand who I mean.
Faculty are divided as well into two large groups: tenure track and non-tenure track. Tenure-track jobs are what most people are familiar with, at least in a vague way. The tenure track was supposed to be (and pre-aughts, was) the ‘standard’ career path for an academic at a university. That’s the system everyone knows, if they know a system. But another system was made. And that brings us to non-tenure track positions, both permanent and temporary, full-time and (fake) part-time (which are often actually full time), which will consume most of this post. We’re going to break these up primarily between full-time non-tenured or teaching track positions and notionally ‘part time’ or adjunct appointments, but there are a few other types thrown in there. Crucially, this other system makes up the majority of university teachers, around 67% and rising.
But before we dive into the range of non-tenure track positions which make up the majority of college professors today, we should talk about the tenure track because, again, this is how the system is supposed to work and also generally how the public imagines the system does work (even though it really doesn’t anymore). So let’s first look at that, how the system is supposed to work.
A tenure-track position begins with a national (or international) search and a fairly long hiring process (form job-posting to job-offer usually takes around 6-8 months). A newly hired professor is an assistant professor, which means they are on the tenure track but do not yet have tenure. Instead, after about five to six years, they’ll go up for tenure review, where a committee of faculty int heir department along with some external reviewers will look at all of the work the professor has done since their appointment and either recommend them for tenure or not; the university leadership structure typically has a role in confirming a grant of tenure but this is generally a rubber-stamp role. By far the most important part of tenure review at large universities is research; this is the part of the system that is ‘publish or perish.’ Untenured tenure-track faculty (so, assistant professors) represent roughly 9% of all faculty members in the United States, according to the AAUP.
A professor that passes tenure review becomes an associate professor, which confers tenure (making it difficult to fire them) as well as a bump in pay. After another few years, they can go up for review again for promotion to the next rank, simply professor (often termed ‘full professor’ for clarity), which comes with another bump in pay. This second transition is different from the first though; whereas the review from assistant to associate professor is an ‘up or out’ moment (you either get tenure and stay or get rejected for tenure and leave the department), some professors can and do remain associate professors forever. Finally, a handful of professors who really distinguish themselves may wind up with an endowed chair and we tend to call these folks distinguished professors, though their actual job title will usually be something like “the so-and-so Professor/Chair of this-and-that” where the ‘so-and-so’ is the name of the donor that endowed the money being used for the distinguished professorship. Tenured professors represent roughly 24% of all university professors according to the AAUP, meaning that the total slice of tenured or tenure-eligable professors in higher education is just 33% – one third.
Let me say that again: only one third of all faculty work the way all of you think all faculty works. Just one third. This is a big part of what I mean when I say that the United States’ university system is being pillaged without the public knowing; if you told most people ‘only one third of college instructors are actually professors, most of your little Johnny’s classes are taught by non-professors now,’ they’d be shocked! But that’s the current situation.
Tenure-track professors generally teach a fixed course-load, expressed in most cases as a load over semesters, so a “2/2” (pronounced ‘two-two’) load is four courses a year (two in each semester). Tenure-track faculties at research-focused universities (which are all of the flagship state schools) generally teach a 2/2 load; mixed research/teaching schools (your third-string state schools and less well-funded private schools) often have 3/3 loads. Teaching-focused institutions may have 4/4 or 5/5 teaching loads (or more) and of course fractional loads (like a 2/3 etc.) do exist, but are less common.
In addition to teaching, tenure-track faculty are expected to publish research and do ‘service.’ We’ll talk in another post more about these demands (indeed, we’ve talked about research already), but they deserve a few words here. The amount of research demanded varies by the level of institution; at an R1 the general expectation for a faculty member going for tenure in a humanities department is that their book is out and they have a good number of articles and other publications besides. At less research-focused universities, you might see instead that tenure is set at a certain number of articles and the book is instead at the jump to full professor.
Meanwhile ‘service’ refers to all of the non-teaching roles faculty fill in a department. The university is predicated on self-governing departments of academics (‘colleges’ in the literal sense of an association of colleagues) and so departments are effectively run by committees and faculty appointed to do various key roles: student advising, graduate admissions committees, hiring committees, committees on teaching, and of course department chair (and possibly vice or assistant chairs) who steers the department. Of course faculty are assisted in those roles by the department staff who handle much of the paperwork, compliance and book-keeping. Some, but by no means all, of these service jobs come with a ‘course release’ which is to say the faculty member teaches less in order to do the extra service, but there is an expectation of a certain amount of service work always being part of the workload mix.
Finally, the more important service positions are often restricted to either associate or full professors – you have to get tenure first before you get a particularly loud voice in the running of the department. Nevertheless, even assistant professors are going to be ‘in the room’ when decisions about courses, resource allocation, scheduling, and so on are made, which matters quite a lot. Moreover, because even assistant professors are expected to become permanent members of the department, their interests tend to be considered because, well, frankly, the tenured professors have to live with them for the next few decades, so you might as well be friends. This fact is really important for understanding why departments can be so callous to anyone not on the tenure-track (and why tenure-track faculty can be so oblivious to how callous they are being), because NTT faculty are usually not in the room when decisions are made.
Which brings us to:
We should start our look at the range of NTT teaching positions by again breaking these down into categories. Unlike the tenure-track, where there is a clear progression of positions each with a standard title, NTT positions are a confused jumble, often by design, with very different positions often sharing job titles. A ‘teaching assistant professor,’ for instance, may well be a permanent member of the department, or someone on a five-year non-renewable appointment, or someone around for just a single semester, paid by-the-course as an ‘outside’ contractor. One is left to strongly suspect in many cases that this confusion is intentional, with universities and departments using job titles as a means to obfuscate just how much of their teachers are not permanent faculty (note, for instance that almost no one advertises jobs with the word ‘adjunct’ in the job title anymore).
So instead I want to break down these positions by conditions of employment. On that basis, we can break down appointments into four basic types. There are permanent, salaried non-tenured full-time teaching positions which we’ll call teaching track faculty. Then there are non-permanent but full-time long-term non-renewable versions of these positions which we’ll call visiting assistant professors or VAPs (though the terminology around them is variable). Next there is the rarest bird in this category, professors of practice, where a professional in a field also teaches that field part time for a university. Finally, there are short-term ‘part-time’ positions, which we’ll refer to as adjunct appointments.
We can start our look at non-tenure track professors with professors of practice, generally the rarest sort of NTT faculty and also the one that universities would like to talk about the most. A professorship of practice is generally a non-tenure-track appointment created for individuals successful in the non-academic field so that they can teach in in that field (often despite lacking the normally required degree, like a PhD). So for instance, a civil engineer might also teach part time as a professor of practice or do so after retirement. As the American Association of University Professors notes, professors of practice are the most likely of all NTT professors to have terms of employment (pay, benefits, teaching load) which approximate the conditions of tenure-track faculty (but without tenure or generally a strong or meaningful voice in the running of their department). In particular, professors of practice often have long-term contracts (say, 5-years) which are presumptively renewable, in contrast to much shorter term contracts for most other sorts of NTT faculty.
That said, the big thing to know about these sorts of faculty is that while universities love to present the typical adjunct as this sort of thing, the practicing dentist teaching a course or two on the side at the local dental school, in practice they are a tiny minority of professors, probably much less than 10% in fields where they are common and almost entirely absent in many fields (like history, for instance).
Substantially more common are effectively permanent ‘teaching track’ (also sometimes called ‘professional’ track) faculty. Because a common title for these positions is teaching assistant professor they are sometimes collectively called ‘TAPs’ (matching the VAPs below). Teaching track faculty generally aren’t eligible for tenure, generally get paid less than their tenure-track faculty (but are paid on a full time, salaried basis, separating them from adjuncts; at some institutions they come quite close to salaries of tenure-track faculty, at others they might be paid around half as much), generally teach more courses and typically do not play a meaningful role in the governance of their department (since those roles are largely reserved for tenured or at least tenure-line faculty), though they may be expected to do some kind of departmental service. Unlike professors of practice, teaching track faculty today almost always have PhDs in their field; the days in which this sort of appointment could be obtained by someone with an MA are effectively over (and indeed, have been for about two decades). The thing that defines these positions collectively is that they are full-time but non-tenure-track.
These positions often go by a bewildering set of names. Perhaps the most common is to take the traditional assistant/associate/full professor ladder and attach the word ‘teaching’ to the front of them to make ‘teaching assistant/associate/full professor,’ but as that phrasing has become more common, it also gets used to paper over what are clearly adjunct appointments. Likewise, teaching assistant professors are sometimes ‘disguised’ as professors of practice in their job titles (leading to the curiosity of ‘professors of practice’ whose ‘practice’ is ‘having a PhD in their field and a traditional academic background).
I should note, because I’ve seen students (and regular people) befuddled by this before, but when I say that teaching track faculty are not eligible for tenure, I really mean not eligible under effectively any circumstances. Because tenure-track searches are functionally always external and because it is (and this is going to be a trend) extremely rare to consider internal candidates seriously in those searches, a teaching track faculty member’s contribution to a department isn’t going to matter because that department is extremely unlikely to consider them for a TT hire. This is compounded by the fact that at large universities the culture of the tenure track faculty strongly holds that tenure-line decisions are based on research and not on teaching, so even for another department, achievements in teaching are unlikely to matter very much. Consequently, there is functionally nothing a teaching track faculty member can do within the scope of their actual job duties to try to move from one track to the other. Indeed, even spectacular performance, things like winning the student-voted best teacher award three years out of four, for the entire university, won’t do it. I have never once ever heard of a department hiring a teaching-track faculty member to the tenure track for any reason, from teaching to scholarly excellence. I’m sure it has happened somewhere, when the planets were aligned under a blue moon, but it is rare in the sense of ‘most departments will never do this once.’
And we should also note here visiting assistant professorships, also known as VAPs. In practice, VAPs generally work like a time-limited form of a teaching-track appointment in terms of the conditions of employment, but they are often held by early career scholars who are still on the job market, whereas faculty with permanent teaching track appointments have often exited the job market and intend to stay long-term where they are.
Once again, the terminology here is tricky; what I mean by VAPs in this category are term-limited, full-time appointments (so, say, ‘full time for two/three/five years without an expectation of renewal,’ though some VAPs might be renewed). On the one hand, many positions with VAP as the job title are actually adjunct positions (discussed below). On the other hand it is also frequent in the humanities for many post-docs (‘postdoctoral research fellow/associate’) to actually be VAPs in disguise. You can tell because the idea of a post-doc is that it is supposed to involve relatively little teaching and lots of research, which is why the word ‘research’ is in the full name, but it is now common to see ‘post-docs’ that involve full (2/2 or 3/3) teaching loads, at which point they’re hardly post-docs; they are just VAPs with a fancy name. Meanwhile I have also seen a trend for second-tier institutions (which may or may not be phasing out tenure) to ‘trim’ the ‘V’ off of a VAP, calling it an ‘assistant professorship’ – a lie exposed as soon as you see ‘non-tenure track’ or ‘non-renewable’ (or both) in the job posting.
Note that actual post-doctoral research fellowships are far, far more common in the STEM fields than in the humanities. We’re not going to deal much with that system here, but in brief, in many STEM fields, time as a post-doc researcher is effectively required before one can get on the tenure-track. Post-docs of this sort thus in theory are a kind of apprenticeship system, although my understanding is that the expectation here is that this ‘apprentice’ stage involves a lot of winnowing and burn out. By contrast in the humanities actual research post-docs mostly serve as gilded lily-pads for PhD students coming out of elite institutions, enabling them to burnish their CV while staying on the job market; there aren’t anything close to enough of these sorts of post-docs – indeed, even if one includes ‘teaching’ post-docs, there are not enough – in the humanities for a meaningful fraction of even PhDs of the top ten programs to go through one. Such post-docs in the humanities are actually more selective than tenure-track jobs (and indeed, I have come far closer to landing the latter than I have ever come to being even seriously considered for the former).
So to recap, you have permanent full-time teaching appointments (teaching track) and temporary full-time teaching appointments (VAPs), along with professors of practice, making up the normal full-time non-tenure-track appointments. Collectively, these full time non-tenure-track positions make up about 20% of all faculty appointments and their percentage has been rising over time. In particular these kinds of appointments tend to be common at the top-tier of universities: R1 (top-level research) universities are generally 50% tenure-track, 23% non-tenure-track and 27% adjunct, whereas colleges and universities offering only master’s degrees (so we’re moving down the university funding ladder) are 32% tenure-track, 15% non-tenure-track and 54% adjunct, while colleges that only offer associates degrees are 18% tenure-track, 17% non-tenure-track and 65% adjunct. As you can see, as one marches down the university prestige ladder, both tenure-track and teaching-track fade to an ever larger and larger share of adjuncts.
And all of that at last leads us to:
It is by this point quite rare, actually, for most universities to include the word ‘adjunct’ in a job title; it used to be much more common. But as the adjunctification of academia became a real and visible problem, universities have responded not by addressing the problem, but by disguising it. Consequently adjunct appointments have a bewildering array of names and titles which in practice in my experience make functionally no different in terms of the kind of appointment.
So for our purposes, an adjunct appointment is a ‘part time’ limited term teaching appointment. In particular what makes adjunct appointments different is that adjuncts teach on short-term contracts which pay them per-course taught, like an outside contractor, rather than a salary. This arrangement is convenient for universities because it means adjuncts do not need to be fired, they can merely be not-renewed, a point that came up in the recent Hamline controversy. It is also convenient for departments because it allows them to trim their adjunct work force as necessary to the particular teaching needs of the moment. In this sense, the adjunctification of higher education is effectively the gig economy, applied to university professors.
Because hiring lots of adjuncts is a practice already in ill-repute, the tendency is to disguise these positions in terms of job title. While ‘adjunct instructor/lecturer/professor’ used to be the common titles, today they are increasingly rare. Instead in my own experience I’ve seen what are clearly adjunct positions described as ‘instructor,’ ‘visiting instructor,’ ‘visiting lecturer,’ ‘teaching assistant professor,’ ‘professor-in-residence,’ ‘visiting assistant professor’ and even some ‘post-doctoral fellows.’ Those titles allow universities to hide their adjuncts among their actual VAPs, TAPs and post-docs
The working conditions for nearly all adjuncts are shamefully bad, which is why universities and departments go to such lengths to disguise the nature of those appointments. While all non-tenure-track academics have limited job security, adjuncts have effectively none, since they need to negotiate new contracts every semester or every academic year. This job security question is an important one because academics are, of course, supposed to talk about difficult subjects and say difficult things; one is left with the strong sense that university leadership prefers adjuncts because they lack the sort of protections that make academic freedom work.
At the same time, adjuncts are paid awfully. As noted, adjuncts aren’t paid a salary but rather contracted on a per-course basis – they are effectively freelancers (and if you are thinking ‘freelance teacher’ sounds like a terrible idea, well, it is) – and the per-course payments are typically extremely low. The average per-course pay is around $3,556, though that conceals a lot of variation, with some adjuncts paid closer to $8-10,000 and many, many more paid less than $2,000 per course offered. At a 2/2 load, an adjunct being paid that way would be paid a total of $14,224 per year, without benefits, compared to a tenure-track professor who might be paid $60-75,000 (in the humanities, more in STEM or business) with benefits to teach the same amount.
Now I want to note something, which is that these appointments are often ‘part time’ in name only. Most universities carefully calculate FTE so that an adjunct can teach as much as their regular faculty while still remaining under the 0.75FTE legal standard for ‘full time.’ In practice, many adjuncts are thus forced to string together multiple different adjunct appointments, or appointments with extremely heavy teaching loads, with each university using the ‘part time’ nature of the work as an excuse not to offer things like family leave or health benefits which would be required by law if an adjunct was a ‘full time’ employee. The result is a system which encourages adjuncts to invest as little time as possible into each class they teach (with deleterious effects to the quality of education), while at the same time relentlessly burning them out. It’s an awful system for student and teacher alike.
Now you may ask why anyone would take a job like that with poor pay (for a job that requires a PhD!), no job security and no benefits. And of course the answer is ‘because they have no other choice;’ leaving academia, even temporarily for a non-academic job is generally a career death sentence, so as the academic job market contracts, it creates a supply of adjuncts looking to stay in the game. That said those adjuncts are looking to stay in the game for hiring at other institutions; just as no department hires their own teaching track faculty for tenure track positions, it is vanishingly rare for any department to hire their own adjuncts for the permanent, tenure-track version of that adjunct’s position. Indeed, while I know several colleagues who have been (verbally) promised this by a department, I do not know anyone who has ever been hired this way.
The proliferation of adjunct instructors is, however, clearly bad for higher education. The higher education model is predicated on the notion of the scholar-teacher who is engaged in at least some level of research (the amount varies by institution, from research heavy R1s to teaching heavy SLACs and community colleges) and teaching on the premise – correct, I would argue – that those two tasks enrich each other. Teaching a topic stimulates research thoughts on it, while a research agenda keeps the teacher up to date and current on the state of knowledge in a field. But an adjunct instructor is not paid to do any research and may well not have the time to do so.
(A convenient time to remind you all that my writing both here and also my research writing, is paid for by…you, dear reader, should you opt to support me on Patreon.)
Moreover, most adjuncts in order to make ends meet have to stack multiple heavy course-loads due to the shamefully low pay they receive, and so while many adjuncts are dedicated teachers they are rarely able to give each class the time it needs. That is compounded by the fact that the short-term nature of adjuncts means they have little freedom in what they teach, since getting a new course ‘on the book’ takes time and is thus impossible for an adjunct with short-term appointments. I have been repeatedly asked by students when I would teach a course on Greek or Roman warfare and the answer is ‘never’ despite tremendous student demand because I am never in an appointment long enough to propose and get approval for the course to be on the catalog, as opposed to tenure-track faculty who generally have far, far greater freedom to shape their course offerings.
Consequently, adjunctification is a blight on academia, reducing the quality of research and teaching our universities produce, degrading the student experience and betraying the fundamental reason why the public funds these institutions in the first place. So it will be no surprise that it is a growing phenomenon. In 1985, TT-faculty made up a simple majority – 53% – of all faculty appointments, while adjuncts made up only 33%, with that 33% frequently consisting of instructors without PhDs or PhDs quickly transitioning to tenure-line jobs. Today, TT-faculty make up just 33% of all faculty, while adjuncts make up 48%. Adjuncts are by far the most common type of university ‘professor,’ more than doubling the next largest category (tenured professors at 24%).
As noted above, the slice of the faculty that are adjuncts varies by type of institution, from around 30-40% at PhD-granting institutions to above 60% at associate’s colleges. But the fact is, even departments at top-tier R1 publics often rely on adjuncts to fill teaching gaps which should and in the past would have been filled by a tenure-track hire.
So to recap, there are three major types of faculty: tenure-line faculty (including tenured and tenure-track faculty), permanent ‘teaching track’ non-tenured faculty and adjuncts. Of these, the size of the last group has absolutely exploded. The job of ‘professor’ as the public imagines it, has functionally ceased to exist in much of higher education, and where it survives, it is ailing. One thing to note in the chart above is how tenured academics also far outnumber tenure-track academics, as universities cut new tenure hires (replacing them with adjuncts) and just wait for the last tenured professors to retire.
There are some (all too) easy implications folks tend to want to take form this information which I think we first need to dispel. The first of these is that the hiring situation in academia is the result of ‘elite overproduction.’ What I hope you can see in the data above is that it isn’t that the demand for higher education teaching has gone away, but rather than the conditions under which it is done are changing. University leadership have exploited the creation of an academic caste system to create a class of academic serf, allowing them to redirect funding (and spiraling tuition money), often towards their own pet projects. But the total number of teachers you need at this level is not declining. This is not ‘elite overproduction’ but the gig economy run amok in a work environment that used to work much better for both teachers and students (and now works well primarily for university trustees and chancellors).
Second, this is not – or at least has not in the past been – a red-team/blue-team issue. Adjunctificaiton does not, in my own experience and discussions with colleagues, seem to vary meaningfully between red states and blue states. Blue states have been aggressive in cutting public higher education funding just as much as red states and continue to do so. Without a doubt, the assault on tenure in Florida and Texas will make this problem worse but only worse by a degree, which is itself a dreadful statement on the state of academia.
But there are some important implications to talk about here which also speak to the question of ‘what is to be done?’
The first thing to note is that the rise of the adjunctified labor market has served to fairly obviously weaken the positions and protections of the shrinking tenured minority. One reason entire states are now thinking of abolishing tenure (in order to sustain a politically motivated assault on their own higher education systems) is because they know given the shape of the job market nationally that replacing tenured professors with adjuncts or teaching-track faculty will be easy and cheap. Consequently, the scourge of adjunctification negatively impacts the tenured and tenure-track of our disciplines as well.
However, by and large the tenured and tenure-track members of most disciplines, including mine are complicit in the system of adjunctification, despite vocally despising it. This is not a statement I enjoy writing, but I think it is unfortunately true. The issue here is the one thing that the TT-faculty still control in all of this, which is who gets hired, particularly for tenure-line jobs. Remember, hiring is done by a committee of faculty members in a department! One response to adjunctification would have been to cling to solidarity within the field, insisting that adjuncts ought to get full consideration for tenure-line jobs (both in their departments and in other departments), that tenured academics should of course support labor actions by NTT-faculty, and that departments should, as much as possible, refuse to rely on adjunct labor and instead at least insist on hiring permanent teaching-track faculty (and then be willing to tenure-line appoint them if they excel).
At least in my fields (Classics and History) departments have done effectively none of this. Instead, the norm remains a caste system: some lucky PhDs receive tenure-track jobs almost immediately on graduation and never spend any time in the adjunct/teaching-track treadmill, while other, equally capable, academics who miss those early hires are left in an academic underclass where the very fact that they have to work as adjunct or teaching-track makes their own departments as well as others unwilling to give them fair consideration for permanent, tenure-track appointments. And of course no department says they’re doing this, but how else does one explain a hiring system where experience manifestly hurts applicants, as you can see here:
That caste system, whereby one is either anointed a Brahmin or condemned to live a Shudra at academic ‘birth,’ in turn makes it very easy for tenured academics to ignore calls for solidarity with their non-tenure-track ‘colleagues.’ One of the things that was notable, for instance, about the recent (this year!) Rutgers strike was that it was one of the first times ever that tenure-line faculty actually stood in solidarity with striking NTT-faculty or graduate students. I was at UNC as an adjunct for the UNC graduate student strike of December, 2018 and my sense was certainly that the majority of faculty were more concerned for the impact that the strike might have on students getting their grades in a timely manner than they were in the poor working conditions of graduate students.
It is my hope that the Rutgers strike is a vision of the future, that at long last, with the tenure-destroying barbarians at the gates, the tenure-track members of our fields, who have far more power in this system, have realized that their tolerance of an academic caste system has sold university leadership the rope with which it plans to strangle tenure.
So what should be done? I think the crucial first step is to break down the academic caste system by shifting hiring standards; as noted this is one thing TT-faculty control. Instead of hiring ABDs and very recent graduates of elite colleges, TT-faculty should demonstrate that we are one field by focusing hiring on promising scholars currently teaching as adjuncts of teaching-track faculty, placing value on experience and a proven track record of scholarship rather than on pedigree. Departments that fail to do this, quite frankly, should be shamed in their fields. It should be as disreputable for a department to hire a fresh graduate when there are so many more experienced candidates as it is for departments to hire their own graduate students.
Moreover, departments need to offer more than a token resistance to pressure to fill out enrollments by stocking up on adjunct or underpaid teaching-track appointments. I am not, for what it is worth, entirely against the idea of a ‘teaching-track’; some academics really like teaching and rather don’t like research much and there should be a space for them. That said, these positions should still be eligible for tenure and promotion, and in nearly all universities, they are not. Indeed, my own preference would be that they be paid and tenured on the same schedule as traditional ‘research-track’ tenure-line appointments, just with different expectations for achieving tenure (more teaching, more advising, perhaps more service, less research). Is it a risk for departments to refuse to hire underpaid, untenured academic serfs? Absolutely. Doing the right thing is often risky, it brings personal consequences. That’s why we value it so much; professors with tenure who are extremely hard to fire should at least be able to summon this tiny amount of courage. Those who cannot are not worthy of the tenure protections they clearly never intend to use.
Finally, TT-faculty should operate under the presumption that they will support pressure by NTT-faculty and graduate students for improved working conditions. The default position should be support and TT-faculty need to place that position ahead of the supposed need of students not to be troubled overmuch by the exploitation of an academic underclass. The students will be fine, but your NTT-colleagues and graduate students need your help.
And what about for the public? As we’ve noted, the increasing prevalence of adjuncts in higher education has a negative impact on the scholarship and teaching universities produce. Now if private schools want to offer an inferior product, that’s their choice, but there is no reason the public should tolerate the pillaging of public institutions built with taxpayer money. State governments have near total control over public institutions and can exercise it, conditioning funding on the creation of tenure-line appointments to replace adjunct appointments and requiring a higher proportion of university funds be directed into instructional budgets and away from administration or student amenities. This is a result public outcry could produce, one which might also help to curb spiraling tuition costs (and the connected student debt problem) and it should happen.
What the public deserves out of its state-funded institutions of higher education is a faculty of scholar-educators who both push the bounds of human knowledge and communicate their expertise to both students and the broader public. That mission is not possible with precarious, untenured appointments, it is not possible with the largest group of instructors overloaded with teaching at extremely low wages, it is not possible without tenure to protect academics who say unpopular things. It is not lost on me that at the moment the system is moving in the wrong direction, with some states and institutions preparing to abandon tenure entirely and others effectively phasing it out by adjunctifying their entire teaching faculty.
But it does not have to keep moving that way and for both the good of students and the public, it ought not continue moving that way. The public ought to demand that their higher education dollars are used for their intended purpose and that intended purpose includes professors. Not visiting lecturers, not adjunct instructors, not professors of practice who aren’t, colleges of professors who are colleagues of each other rather than arranged in an academic caste system which benefits university leadership and no one else.