Margaret Mitchell on AI
These systems, she said, are terribly suited to being integrated into search engines. “They’re not trained to predict facts,” she told me. “They’re essentially trained to make up things that look like facts.”
– From an Ezra Klein piece on AI in the NYT
Five things I like right now
Super into Ibrahim Maalouf.
Severance. Just started it though. No spoilers!
Just powered through the first three books of Scott Lynch’s Gentlemen Bastards Sequence.
Made Babish’s Ted Lasso shortbread for the second time. Takes all day, but is so worth it.
Technology reveals the irrationality of organisations.
Big week? Huge week.
Something different. Using Mathias Jakobsen’s Question Cards rather than a play-by-play. Constraining these answers to work-related topics.
What do I wish would be different?
It’s gone from “boy it’s quiet” to “gosh it’s busy” in the space of a week. So while busy is always better than quiet in consulting, it all feels a bit frantic.
What am I happy about?
Well, the new busyness is good. And the busyness comes with associated interesting projects which is also good.
Which beliefs are holding me back?
It’s really apparent this year that my role is very much not on project delivery. It’s a bit less clear what my role actually is, other than “leadership” which I’ve never been comfortable with. However, a colleague said to me recently that they thought I could start a cult if I wanted, which I choose to take as an acknowledgement of leader-y-ness.
So it might be time to lean into that discomfort with leadership.
When is my deadline?
My reaction to deadlines is always the Douglas Adams quote but in this case I honestly don’t know what the deadline is.
The point isn’t to win or end. The point is to keep playing.
How to Read an Academic Paper in Ten Minutes
Reading an academic paper in 10 minutes is
a bit of a cheat a clever tactic, but if you really do need to read it, you’re going to have to read the whole thing.
I used to give this advice to honours-level students when I was an academic. A friend prompted me for it yesterday so I dug this out and tidied it up.
In your career as a student, a researcher or a curious person, you will come across many more academic papers than you actually need to read. “Reading” a paper in “10 minutes” is actually about using the structure of a typical academic paper to help you skim it in a constructive way. You should do this to decide if you need to put in the effort to read the whole thing.
The ultimate aim is, of course, to read the smallest number of papers as possible while having read all of the papers you need.
First, read the title. Titles of academic papers are generally bad. Move on to the next step.
Second, read the abstract. The abstract is the authors’ summary of what they think the paper is about. Start with the first sentence, then skip to the last sentence.
If you’re convinced this is completely the wrong paper, STOP. Otherwise, you should probably read the whole abstract.
Third, skip right to the conclusion. Start with the first sentence, then skip to the last sentence. If you’re convinced this is completely the wrong paper, STOP. Otherwise, you should probably read the whole conclusion.
The conclusion will work in one of two ways. One way is that it will tell you the actual results of the research that was written about in the paper. If the results aren’t what you need, you should probably STOP. If they are interesting, keep going.
The other way a conclusion can work is it will say how the results of the research fit into a bigger picture. If you encounter this type of conclusion, you might need to jump back to the section before the conclusion to find the results. Just like the other type of conclusion, if these aren’t the results you need — STOP. Otherwise, keep reading.
If the results seem interesting, you are probably committed to going further, unless you’ve already decided that you’re not going to read this paper at all.
Maybe check the references
If you get this far, but don’t end up reading the whole thing, you might look through the references or bibliography for other, related, papers that are closer to what you’re actually going to read.
Fourth, go back to the beginning of the paper and skim through the headings. This will give you an idea of the structure of the paper.
Fifth, look at the pictures. See if there is information in the figures, tables and diagrams that you can use to decide if this paper is telling you something really useful.
By now you are well into it. You should know what the authors of the paper are trying to tell you, and you should be able to follow someone else talking about this paper in a seminar.
(HINT: do up to this step to “read” the papers for seminars that your peers are giving)
If you are now convinced that this paper is not what you need: STOP. Otherwise, keep going.
Sixth, read the topic sentences of all the sections from the introduction to the conclusion. A topic sentence (in a well written paper) is the first sentence of each paragraph. Perhaps you could read the first and last sentence of each paragraph if you’ve got time to spare. Or if the paper is poorly written.
By now, you have “read” the whole paper. Congratulations. Even more than that, you’ve been through it at least twice so you now know how it is structured, making it even easier for you to find things in it later.
Seventh, go back and read the abstract. (Yes, again!) You do this to cement in your head the authors’ summary of what they think the paper is about. Perhaps you disagree with them? Then you need to decide if they’ve written a bad abstract or if your understanding is flawed. (Or possibly both.)
Now, if you get to this point, you should more than likely STOP. Read no further unless you so completely convinced that this paper will allow you to progress in your own research that it is worth spending the 10-30minutes to (re)-read the whole thing.
Using this method to “read” papers means that in the time it takes you to plod through each and every paper you come across, at 20-30 minutes a paper, or two an hour, you could get through six or more! (Or, let’s be honest here: you could get through the same 2 papers in less than half the time!)
“Reading” a paper in “10 minutes” is
cheating a tactical approach to using the structure of a paper to help you skim it productively.
If you really do need to read the paper, you’re going to have to read the whole thing.
Microsoft has made the unfortunate decision to limit Sydney to make it a better search engine (boring) rather than answer what I think is the more interesting question, which is how do you ethically and responsibly manage a deranged chatbot with knowledge of the entire internet.
– Ryan Broderick, obvs
Weeknote CW06 (S4E4)
Quick and dirty and cryptic:
Monday: it’s funny how often clients fall into the “not enough money to do it right; always enough money to do it twice” situation.
Tuesday: Worldwide we’re a centre of gravity for our parent-company’s parent company in terms of what we do. PC’s PC is enormous. Each time we meet someone we need to explain what we do.
Wednesday: PC’s PC brings the opportunity for para-academic work! Just when I think I’m out they pull me back in.
Thursday: Delivered internal training on research capability.
Friday: how do we do project management internally? And also, spent a long time doing in-person business development. A good day.
What is a “household”?
everyone who is in the room when a cartel draws up a standard definition of what constitutes a household is almost certainly drawn from a pool that is more likely to have a summer villa than a child doing domestic work or construction labor half a world away. These weirdos, so dissimilar from the global majority, get to define the boxes that computers will shove the rest of the world into. If your family doesn’t look like their family, that’s tough: “Computer says no.”
Cory Doctorow has no time for your bullshit
A few of his fans were unhappy that I said I thought Donaldson was an algorthimically-optimized charisma void that looks like someone who would wear pants that zip off into shorts.
— Ryan Broderick on a certain YouTuber
Weeknote CW04 (S4E2)
As we say at my house, there was too much week in this week!
Personal, home and family life needed to be prioritised so I took leave on Wednesday and Thursday to create some space for it.
Company names are always made up and change week to week.
Monday: Another typical Monday. Reviewed a proposal from late last year for Padlock Microsystems seem to be ready to commit. Continued grinding my way through onboarding for a relationship with a SaaS provider in our space. It’s a lot. Also chatted to someone who’s looking to escape academia and pivot into UX.
Tuesday: A brief business development chat for a fun project for Boarstar who do something fun but have a tiny budget. The sort of projects that we do for a treat in between working for the banks. Ongoing project for Forest Electronics is gathering steam and the team on that are doing great work herding the client cats. Ended the day with a call with Padlock Microsystems who are super keen to start in the coming week. Relied on a jargon-free Cynefin-style replay to them of what they were talking about and I think I won them over.
Wednesday: Home stuff.
Thursday: Exhausted from Wednesday.
Friday: Still exhausted from Wednesday. Spent the afternoon finishing writing up case studies for the new company website.
I'm @email@example.com on Mastodon