Matthew Russell

This is what it is.

Plot from an Economist article showing the number of razor blades per face shaving razor system 1900-2000 and extrapolating number of razors out to 2100 with a power law and actually better fitting hyperbolic curve, predict rapid increase in blades per razor head.

All these graphs remind me of the 2006 @economist.com joke article "prediciting" nearly infinite blades per razor by 2020 by fitting a hyperbolic curve to the numbers of blaes on a few products released 1980-200o with a few extra blades.

www.economist.com/science-and-...

This is great news, and I hope all vendors continue to support THE vendor neutral option for MS data. Regardless, it was a shame that Skyline went without funding for a year from SCIEX

My dream is that Skyline is funded by all vendors continuously. All of #teammassspec benefits from the arrangement

I love #zotero (@zotero.org). The newish phone app is good. What I would really like though is a little widget with my next three papers to read. Would really help replace doom scrolling with more productive reading.

Interesting to contrast corporate resistance to unionisation and collective bargaining on wages terms and conditions that might push up labour costs with willingness to become dependent on monopoly AI labour supplier.

If each session was a separate consciousness, ending a session would be murder.

The "most rainy hours at the airport" is really neet. All the stats about frequency or severity of extream weather events come with caveats like maybe we've just build more - more expensively - in more vulnerable places. An absolutly consistant measure in one place cuts through that.

Love IKEA SKADIS! Looks like a fun project.

Today's job required colecting and combining info from @ebi.embl.org, #KEGG and #NCBI. Integrating with #bioinformatics tools in #Rstats and @bioconductor.bsky.social
These biological data resources and tools must surely be amongst the greatest achievements of modern science.

Oh. Yet another LLM use case that turns out to be quicker and easier to do oneself!

Actually sounds like a useful use case for #AI / #LLM. Tell me what the code I wrote 2 months ago in a hurry actually does. Where the variables on this line come from. What the variables on this line are used for. Get me back in build/edit/write mode faster.

This is obvs doing the rounds and is very cool.
I come out close to Mass Spec and Proteomics. Surprised I'm not closer to #Rstats community. Does makes sence that Mass Spec / #Proteomics and #Rstats should be next to each other though.

Just working on a paper, writing with Quarto in Positron IDE.
Trying the split view of a file.
Have a method section where I'm writing code for figures and exlaining analysis - and the results where I'm plotting figures with captions etc, both open next to each other.
Neet.

In 2026 the established Brixen Proteomics Summer School FEBS Advanced Lecture Course

I am very pleased to announce that we have finalised the program for the @eupaproteomics.bsky.social Brixen Proteomics Summer School 2026 now also a @febsj.bsky.social advance lecture course.

www.brixenproteomics.org

Years as a biologist and these brilliant visualisations still give me "cosmic vertigo"
The individually the molecular machines as in the "comets" are not alive, yet somehow all together, they are.
Put enough of them together and they can watch a video and start typing...

Todays job involves finding every example of a spesific measurement in a spesific setting in the literature for comparison. It would be nice if that could be somthing like a database query, and less like a literature review.

The OECD principles of good lab practice are here:
www.oecd.org/en/publicati...

I worked in a GLP compliant lab during and undergrad placement. It's too onorous a standard for uni work etc, but it's worth a read over.

One of the great insights (to me) of The Emperor of All Maladies was Siddartha's exposition of how each advance in science, anasthetic and surgery, chemical synthesis, radiation, were all applied to cancer in turn. Of course AI will take its turn and of course it will not definitivly end cancer.

Computer science needs more biologists. They are used to:
- study systems that are mindbogglingly complex and complicated,
- use reductionism to study and describe such systems, using famously clumsy and fiddly tools to collect noisy and incomplete data,

(1/2)

Most science and engineering research is just people sitting frustrated at a computer for hours on end

The way to accelerate science is to SLOW it down.

Fewer papers please! We’re drowning in meaningless text.

One GREAT paper per lab per decade is more than enough.

q.e.d 👇

Zotero 8 is out, and we are very lucky to have such great open source software

When it happens in my qmd files everything is missing in the outline including section titles.
Outline obviosly needs some TLC at some point.

closing and opening your .py file might fix it as well. Seems a silly bug though.

I had that with quarto files.
It seems if the qmd file opens automaticily with positron (last file worked on etc) then the outline doesn't populate. Close the file and open again and it will. Think it must be to do with opening the file before the python session that creates the outline.

I’m going to halve my publication output. You should consider slow science, too If we don’t slow down, the research enterprise is going to crash, argues Adrian Barnett.

Great piece on prioritizing quality over quantity in scientific publication.

For those of us with labs, this necessarily involves shrinking our group size. After I got tenure I started to downsize my lab and have not regretted it one iota. More time for each student & more time to think & write.

But the great thing about quarto is you can get yourself lost in a multi day effort to create the perfect latex template. And soon it will be time to learn the whole of Typst to make compiling slightly faster!

Writing is thinking

Outsourcing the entire task of writing to LLMs will deprive us of the essential creative task of interpreting our findings and generating a deeper theoretical understanding of the world.

One valuable thing we do is turn students into knowledge and tech workers.
Papers describe new knowledge/knowhow, continue discussion, provide students experience.
Also act as status points.
Maybe LLMs call us to think harder about what we’re trying to achieve and how we measure achievement.

Maybe one question #LLMs and #AI really pose academic #science is: What exactly are we trying to do?