Coffeehouse notes - June

Underpaid PhD students

probably just whining

Unbelievable how many times I heard people say a degree or PhD is "a chance" so it is okay to be underpaid! What utter nonsense. Any decent job is a chance, since you can put it on your CV and develop your professional skills. Stop with the Stockholm syndrome already.
It is unbelievable that PhD students and, in some places, technicians cannot afford a lower middle-class life without the support of their parents or lucky top-up scholarships. And that is just the beginning of it, undergrads in science and many postgrads have it horrible as well. A first step would be to unionize, but all this will do is stop further deterioration and provide some modest benefits. Nonetheless a PhD student union is essential.
Even if you can make bank after your PhD, the costs of tuition leading up to the degree and the abysmal pay have many downstream effects:

  • lack of equal opportunity is unfair towards underprivileged students and thus unethical
  • the winner takes it all system punishes the unlucky ones, e.g. imagine you have to drop out due to illness or burnout midway through a degree (tuition should be replaced by a real success-taxes)

  • the quality of the average PhD, professor and scientific article suffers because:
  • smart people drop out as they cannot afford tuition, cost of living and the opportunity costs of not working
  • smart people who want to support a family go to industry
  • smart people who are egocentric go to industry
  • smart people who want a comfortable life go to industry

Lifespan testing - what is slowed aging?
This idea developed based on a twitter discussion I had, where someone was doubting the use of mean lifespan differences to measure effects on aging. Finally, getting a chance to dig up references and back up my intuition. Testing some sort of mean lifespan effects remains the gold standard, at least in mouse lifespan studies. The reason is both maximum lifespan testing as well as mortality doubling time require larger sample sizes and are too sensitive to outliers (see e.g. Lombard and Miller 2014). 

Lifespan testing - best practices
Currently reviewing lifespan studies and it is no secret that these are of rather variable quality. In some 80% of the studies I reviewed, the authors forgot at least one of the basics: reporting the strain, gender, and weight of your mice; the sample size, as well as median or mean and max lifespan in a table. Almost none provide the full mortality data underlying the survival curve which can be used for estimating mortality doubling time and other parameters.

Dirty money - clean science?
I was alerted to a new player who will be funding aging research. The Hevolution foundation which is supported by the Saudi royal family. It seems they have been already heavily involved in funding the TAME trial. Many aging researchers will start asking themselves the question whether it is ethical to take their money now. As for me, I have been thinking about similar issues for many years already - due to my exposure to SE/E-Asia so maybe you will find my thinking insightful.

Most countries in the world are not democracies and it makes no sense to sanction them all the time and refuse cooperation. This is not how you survive in a globalized world. Free trade, travel and the exchange of ideas benefit both the poor/oppressed people living in said countries as well as your own citizens. Sanctions, refusal to cooperate and the threat of resignation will be sharper weapons when they are used sparingly and only in response to acute changes, i.e. rate and direction of change, rather than absolute levels of "human rights" or "freedom".

A redditor discussing the news summed it up perfectly with this witty quote:

Behavioral geneticist David T. Lykken wrote:

"If you can find me some rich villains that want to contribute to my research - Khaddaffi, the Mafia, whoever - the worse they are, the better I'll like it. I'm doing a social good by taking their money... Any money of theirs that I spend in a legitimate and honorable way, they can't spend in a dishonorable way"

Chinchilla law of scaling

what a time to be alive

As a disclaimer, I am not in any way knowledgeable about ML/AI, but I do like to follow the news and the general trends because they are relevant to me. By training, I am a biogerontologist, and in my spare time a bit of an armchair futurist and progress "researcher". Alphafold has certainly shown that ML is more than just vaporware. ML is coming to biology. AGI is inescapable (when no one knows).

Many moons ago, when I took a look at the GPT-3 training dataset I immediately thought to myself - would it be better if it trained on more data? Could libgen and sci-hub provide advantages to a company or actor willing to use pirated material? When I asked someone (who I presume was knowledgeable), though, I was told that it is not as simple as that. The systems are not limited by the size of the corpus.

Turns out they are. Apparently these models are quite over-parameterized and it is more compute-efficient to train them on larger datasets with more tokens rather than further increasing model size. It is pretty cool to see how much low-hanging fruit there is left in the ML field.

Anyone got access to DALL-E2

Masking at conferences - yes or no?

quality vs quantity of life debate incoming

I am not really keen to engage the specifics of this debate as it will depend on the country and the conference, however, I do want to make a broader point. COVID has slowly morphed into a War on Extroverts. I know it is incomprehensible for most scientists, since science attracts hardcore introverts, that one could suffer due to social restrictions even as "mild" as "having to wear a mask for three years" -- which pre-pandemic no one would have considered "mild" and that kind of imagery was only reserved for art depicting a dystopian future. It is hard for me to comprehend that it would be hard to comprehend for someone that normal people clamor for normalcy.

Mental suffering is real suffering. Just like depression is not all in your head, neither is social isolation due to the current situation, which is still not even close to normal.

Either way, the masking war will be lost sooner or later since the population at large will not put up with this. (And, no, I am not against masks. I wore one when it still got you strange looks on the bus and wrote articles about the evidence base for masking) We have to find other and better ways to protect everyone involved.

However, apart from that I want to make a philosophical remark here. We - the scientists - promised that the trillions spent on science will buy us a better future, which is why we got that money. No one wants to live in a dystopian future where the air is unfit to breathe. I cannot work as an aging researcher under these circumstances and neither should you. Why extend a life whose quality deteriorates every year? Are we that impotent?

The point is not that we refuse to save lives. The point is that BOTH quality and quantity of life matter. 

Obviously putting on your mask sporadically will do nothing to prevent COVID. It has to be worn consistently throughout social interactions to do anything and that will diminish your quality of life. If someone disagrees with this, they are a bad-faith actor or troll and not worth time to interact with. Okay, having established that: yes, we should wear masks and we have done so for almost two years. But we cannot do so forever and we only do it as long as the risk/benefit ratio merits it. Extroverts have sacrificed plenty during the pandemic.

Why are you doing science?

It's all about the Pinkerian promise of a better life.

Generalizations and stereotypes can be useful, as long as we use them productively and not to judge the individual. This entry is another example of how scientists seem to be disconnected from the life of normal people. 

I always wondered why I do not get along with many scientists. Of course, I have met many cynical people who had no passion for science. It would be understandable if I did not get along with that kind of person -- because I do love my science. However, why is it so hard for me to connect with the workaholics churning out high-impact paper after high-impact paper? Should we not have a common topic and language? After plenty of get-togethers, wine outings, etc I had a realization of what bothers me and perhaps impairs my ability. It is something that I sense is absent in science and scientists: a passion for life, zest, verve, esprit. Perhaps understandably science favors the bookish introvert, and there is nothing wrong with this per se, unless these scientists lose touch with the reality of normal people.

This little dialogue encapsulated what I felt for a long time:
Me: "Why do you do science?"
Them: "Oh, I really enjoy understanding things and playing around. It is my nature."
Me: "Okay, I see. As for me, I do science mainly to improve the life of the people and to better society."
Them: "No, that is not all why I do science."

Okay, maybe this is why you are doing science. However, this is not the reason why you are permitted to do science. A well-run mouse lifespan study costs half a million dollars. A decent microscope costs half a million dollars. This is more than anyone living a comfortable lower-middle-class life in Austria will ever have on their account. This is money that could have been spent to better the lives of those who cannot afford the basic necessities of life. Do you think we - the people - give you money to play around while we clean your toilets, bag your groceries and drive you to your appointments, flights and meetings?

You promised us a better future. We - the people - will destroy you and your science if you do not deliver, whether you think this is fair or not is irrelevant. That is how the world works. Science has to deliver and science has to communicate what it has achieved.

If you prefer, though, there is a more poetic way of looking at it. Science is the only glimmer of hope on a vast canvas of darkness. It is the only way to realize the Pinkerian promise of a better future. The only thing that sets us apart from the dust, the dirt, the cockroaches and worms crawling beneath us.

If you prefer, there is also a cynical way of looking at it. Doing good science can, to some extent, also benefit yourself.

References and notes

Miller, Richard A., and David B. Lombard. "Aging, disease, and longevity in mice." Annual Review of Gerontology and Geriatrics 34.1 (2014): 93-138.