On 18 June I delivered a keynote address to the ANZREG 2024 conference. This was a pretty amazing privilege - talking for nearly an hour to a group of Australian and Kiwi systems and technical services librarians.

A huge thanks to Alissa McCulloch who provided invaluable advice about the earlier drafts of this talk–though of course responsibility for any errors is entirely mine.

Hello. I live in Naarm/Melbourne, just a short walk from the Birrarung where it curves around in a series of turns, folding and looping back on itself. Even though I live in one of the city's oldest inner suburbs, the river is surprisingly peaceful here. In the summer herons dry their wings in the sun and lorikeets shriek as they zoom through the river redgums. On winter mornings like today, a low fog often hangs over the water. The place talks to me about the Wurundjeri people who have lived on this river and this land with the mists and the birds and the redgums for tens of thousands of years.

Traditional Wurundjeri culture, like the culture of other Indigenous people in what is now called Australia, is inseparable from the maintenance of knowledge and learning. Some was embodied, like making a canoe, a coolamon or a net. Some was more cerebral - passed on through stories and mnemonics carved on everyday or sacred objects. And some was experiential, like star maps and Ceremony.

So I want to pay respect to the knowledge and the wisdom and the elders of the Wurundjeri, and Indigenous peoples on all the lands we are gathered on today - to those who lived here many years ago, to those who are passing on their knowledge today, and to the next generation of elders who are emerging.

When Ella asked me to deliver a keynote address at this ANZREG conference, I was quite startled. It's a great honour. Ella told me that I could talk about "Anything–because you have opinions". I think we need more opinions in librarianship. I'd love to provoke you to share some of yours. Hopefully we'll have time at the end of this session to hear from some of you as well.

So friends today we're going to think about libraries, and systems, and library systems.

The description in the program said I'd be using six of my favourite books, but I don't actually have enough time for that, so we'll talk about five instead. Funnily enough, time is one of the things I want to talk about. We'll come back to these themes at the end:

• Lots of things can be true at the same time
• Lots of times can be true at the same thing
• Know where you are
• Share what you know

Sand Talk

Tyson Yunkaporta is an Aboriginal scholar and founded the Indigenous Knowledge Systems Lab at Deakin University. His book Sand Talk changed how I see the world, so I always want to tell people about it. But the thing about Sand Talk is that it's a book about Indigenous thinking that helps us to think through Indigenous thinking. The whole thing is a series of interconnected stories. This makes it very difficult for me to just pick out a nice quote that summarises everything, which I suspect would make Tyson happy.

So instead I'm going to keep coming back to some of the concepts he shares, as we explore the other books. But briefly, for Yunkaporta everything is about relationships and flows. We need to seek out different people and experiences, and learn what we can. In our interactions the knowledge, energy and resources should flow through us and out to others. And all of these connections and interactions should transform us, not just transform the things and people we interact with and act upon.

In Sand Talk there is also a discussion of time. In Indigenous understanding and many Indigenous languages, something can be both a time and a place. A time can be both future and past. This is part of what is sometimes called the Dreaming, though I personally find it easier to think about it using another term: everywhen. We'll come back to this a few times but let's move on to our next book.

Seeing like a state

James C Scott is an American political scientist and anthropologist. His book Seeing like a state explores how states try to enforce "legibility" on their subjects and in the process ignore or override local and contextual knowledge.

What he means is that for big states and empires, it's impossible to govern a large group of people or a large area of land unless you can understand how the people are behaving and what is happening on the land. States need to "see" the people and the land. But it's too hard to govern if you have to understand all of the details about every person and every tree. So large states simplify things so that they can see what they need to see in order to maintain control. That's what he means by making things "legible".

When this is combined with large scale "high modernist" projects like building cities from scratch, or nation-wide agricultural reforms, the conflict between the clarity of the plan and the messiness of the real world usually results in catastrophic failure.

There's a lot packed into 360 pages, but I want to read along with a particular section for a moment, because it gets to the heart of Scott's argument and has a lot of relevance for our own work. A note here–when Scott writes about "bracketing contingency", what he means is assuming that anything the creators haven't considered is basically irrelevant and won't impact on their plan.

Scott writes in part of his conclusion:

The power and precision of high-modernist schemes depended not only on bracketing contingency but also on standardizing the subjects of development.. such subjects... have no gender, no tastes, no history, no values, no opinions or original ideas, no traditions, and no distinctive personalities to contribute... The lack of context and particularity is not an oversight: it is the necessary first premise of any large-scale planning exercise.

James C Scott is writing here about huge state projects like the construction of the city of Brasilia, and the vast and catastrophic changes to agriculture and social life in Stalinist Russia. But I think there are lessons here for those of us working in libraries today as well.

I see most theory in librarianship, and much of its practice, as fitting squarely within the high modernism that Scott writes about. After all–what is classification, cataloguing, and ontological mapping really all about, if not an attempt to render the messiness of the world "legible", as Scott puts it? Every controlled vocabulary flattens reality into a list of predetermined categories and definitions. To an extent this is inescapable, but we should at the very least be more mindful about what perspectives we are including and we're excluding.

Most weeks I think about how absurd it is that the overwhelming majority of libraries of all types in Australia and New Zealand use Library of Congress Subject Headings as the main, or only, controlled vocabulary in our systems. As if the interests and needs of the United States government are somehow universal and relevant enough that we don't need to bother thinking about what local people, local concepts, and local worldviews need in and from the collections we manage and the systems we use to tell the world about them.

We all know how unhelpful it is to shelve physical items based on a classification system designed by an American racist sexual predator from the nineteenth century. Yet the Dewey Decimal system is the one thing most people think they know about libraries.

Even when we "genrefy" collections, you can usually scratch the surface and see DDC underneath.

And whilst there are discipline-specific thesauri and classification systems, and there's been a lot of work happening in the last few years to add things like AIATSIS headings and Homosaurus terms, it would be very rare to find an original catalogue record in a general-collection library that was built on those from scratch and doesn't use LCSH at all.

And this kind of flattening vision happens all the time in other parts of our systems. In recent months I've gone back and forth with a vendor in an ever more frustrating series of support tickets. Needless to say, support, such as it is, is offered in the middle of the Australian night.

Some of our students haven't been able to sign in to Lexis Nexis using single sign on. After some troubleshooting, it became clear that the problem was that the students in question all have only one name each, and the Lexis Nexis platform assumes that they have two different names: a "first name" and a "surname". When we pointed out this flaw in their system, Lexis Nexis insisted that we need to change things at our end so that we sent them two names for each user.

Lexis Nexis sees like a state. If users don't have surnames, we're expected to create one for them. James C Scott tells us in his book that the Spanish colonial powers did exactly the same thing in the Philippines, giving each regional governor a couple of pages from an alphabetically ordered list of approved names. They then went from village to village and assigned family names from the list so that the government could identify people more effectively. Here we are in 2024 expected to do the same thing.

Our support ticket with Lexis Nexis remains open.

This tendency to only see certain things was obvious when the State Library of Queensland launched their Virtual Veteran project just before Anzac Day this year. The Virtual Veteran is a generative-AI chatbot trained on a selection of texts the library holds about the First World War.

Unfortunately the project team at SLQ had a hard time imagining the sort of people I immediately thought of when I first found out about it. Though to be fair, that's probably because I first found out about it not from the library, but from someone who had rather different ideas about how the Virtual Veteran might be used:

The Queensland Government has spent money on an LLM chatbot "trained" on World War I lore in order to "celebrate" ANZAC Day. Please note that it has specifically been given guardrails to not respond to questions about Ben Roberts-Smith, do not under any circumstances try to get around them for comedic effect:

I'm sure you can imagine the sorts of questions this chat bot was responding to pretty soon afterwards.

In a webinar the next week, one of the people involved in building the Virtual Veteran chatbot confirmed that no testing had been done to check whether the guardrails put in place would work. It simply hadn't occurred to them that anyone would ask the Virtual Veteran about anything other than ...well, I don't know. How brave the soldiers had been, or what they thought about mateship? Who did SLQ have in mind as the people who might ask questions of the Virtual Veteran? Maybe they were people like Scott's "human subjects", people with:

no gender, no tastes, no history, no values, no opinions or original ideas, no traditions, and no distinctive personalities.

Well, it turns out people do have tastes, values, and original ideas:

Despite claiming that Charlie has "the persona of a World War I soldier", his responses have the blandly safe tone you might expect from an official government spokesperson. The bot was trained on the 12-volume Official History of Australia in the War 1914-1918; on newspaper articles from the time that were subject to government censorship; and on war correspondence donated to the library. Generative AI by its nature provides confident responses to any prompt based on the content it was trained on. When it was trained on the official narrative of a government that declared war in support of the British Empire, it's hardly surprising that the result is a sanitised imperial view of the world.

What does Charlie have to say to Australians of Turkish, German, or Arab descent? How does the chat bot provide insight into the multiplicity of experiences Australians had during the first world war?

I think we can do better than this.

I can see what the librarians at SLQ were probably trying to do. The Virtual Veteran was an interesting idea that wanted to help Queenslanders engage with their history in a more interactive and intuitive way. Popping down to the library to read all twelve of Bean's volumes about the war in an uncomfortable chair isn't for everyone.

It might have worked with a different framing. The Library could have been explicit about the absences and biases in the collections their chatbot drew from. There could maybe have been different personas to answer different types of questions. Or even better: different personas to answer every question, each having a different perspective.

Because there are always different perspectives. The one thing every historian agrees on is that history is contested.

The Corley Explorer is an example of taking the opposite approach to inviting a community to engage with a library collection. Interestingly, this was also a State Library of Queensland project. The Corley Explorer invites a multitude of overlapping experiences and histories into the library, contributing to the stories the library knows about its collections.

Frank and Eunice Corley drove around the suburban streets of 1960s and 70s Queensland in their pink Cadillac. They took photos of houses, developed them, and then went door to door to sell them to the homeowners.

When 61,000 photographs from the Corleys' basement were donated to SLQ, the library did something very interesting. The Library didn't have much meaningful metadata for this large collection, so it launched the Corley Explorer, inviting Queenslanders to browse the collection and fill in the missing details: geo-locating each house, providing information about what has changed since the photograph was taken, and sharing family histories about who had lived in the house and when.

The Corley Collection could have been a pretty tedious set of broadly similar Queenslander houses from half a century ago, and it wouldn't have said much other than perhaps something about architectural history. Instead of that, SLQ now shares thousands of individual stories, finds connections between them and maintains an ever-growing text for future generations to learn from.

So the Virtual Veteran and the Corley Explorer–two projects from the same library–give us an example of what James C Scott calls "seeing like a state", and also an example of how we might avoid it.

In his book, Scott describes something I think is really important for us to think about as librarians. He writes:

[High modernism's] simplifying fiction is that, for any activity or process that comes under its scrutiny, there is only one thing going on.

I'd love to see more discovery tools like the Corley Explorer. And sure, you could describe it as a "crowd sourcing" project, but I think that really underestimates the work the Corley Explorer is doing. With the Corley Explorer. SLQ has the courage to admit that the collection is full of unknown unknowns, and it uses the Library's ignorance to create connections and encourage storytelling.

The project explicitly encourages multiple perspectives, understandings and histories to layer across each other and connect with each other. It explores memory, community, and meaning. Because of this, every photo is also a mnemonic object like the ones I mentioned at the beginning, prompting memories of childhoods, families and neighbours, and then wrapping those memories into something the library can record.

The Corley Explorer encourages the diversity and human connections that Tyson Yunkaporta highlights as so important in Sand Talk. But it also does the other thing he highlights as important, that we always find so difficult in big institutions: the Corley Explorer changes the State Library of Queensland and the Corley Collection itself, every time someone interacts with it.

Spam

Finn Brunton's Spam: a shadow history of the Internet gives us some helpful ways to understand the weirder and more annoying things we encounter online. Brunton is now Professor of Science and Technology Studies at UC Davis, and studies histories of technology and hacking culture. He has a pretty capacious definition of spam:

Spam is the use of information technology infrastructure to exploit existing aggregations of human attention.

The book was published in 2013, but here's a site someone put together to show how we experience the web in the 2020s, which really helps us visualise what he was talking about:

https://how-i-experience-web-today.com

So every commercial website is mostly spam. But let's touch on some weirder examples before getting into things more specific to libraries.

For something slightly closer to libraries, we can observe the Amazon ebook store. This problem is so old, Finn Brunton flagged it as an interesting new development in his book eleven years ago. But since then, it's expanded into a much more serious problem.

Last August, author Jane Friedman published a blog post called I Would Rather See My Books Get Pirated Than This. What on earth could be so bad that a published author would prefer her books to be pirated?

AI-generated ebooks listing her as the author, being sold on Amazon.

Friedman found at least five titles attributed to her that were obviously AI-generated. Needless to say, she hasn't received any royalties for these books.That's pretty bad, but what happened next shows how optimised and automated metadata systems can really go wrong.

Amazon owns the Goodreads platform used by readers and authors to track reading and talk about books. Friedman found out about these fake Amazon titles because they were automatically linked to her Goodreads profile. Nobody at Amazon knew these titles weren't published by her, because the system was so "optimised" that there were no humans involved at all. And Amazon's response when she complained?

"Please provide us with trademark registration numbers".

Demanding that authors go through the bureaucratic work of trademarking their own names before they take any action against robo-impersonation might seem callous, greedy, and irresponsible. But that's only because it is callous, greedy, and irresponsible.

I quoted James C Scott earlier when we were talking about his book:

[High modernism's] simplifying fiction is that, for any activity or process that comes under its scrutiny, there is only one thing going on.

Well, Amazon insists there is only one thing going on in their ebook store, because for them there is only one thing going on: making gigantic profits. Anything else is just a price worth paying.

Especially if it's someone else who pays.

I reckon Tyson Yunkaporta would identify this as a bad energy flow.

August last year was a pretty busy time for stories about the AI-generated catastrophe that is the Amazon ebook store. Another story showed how high this price could be for those unfortunate enough to purchase a book from Amazon. Fake books by real authors is one thing, but fake books by fake authors can be deadly.

In late August a bunch of AI-generated mushroom-foraging books appeared in the store. One of these was The Ultimate Mushrooms Book Field Guide Of The Southwest: An essential field guide to foraging edible and non-edible mushrooms outdoors and indoors. Perhaps the idea of foraging "indoors" for "non-edible" mushrooms should have been a clear giveaway that something wasn't quite right here.

Some people seek out mushrooms for their hallucinogenic effect, but I think we can hopefully all agree that if you're foraging for mushrooms, you don't want to trust a generative AI that is hallucinating.

Samantha Cole from 404 Media wrote:

False Morels and Death Caps are two species found in the American Southwest that look a lot like their edible, non-poisonous counterparts and can kill you within hours. Foraging safely for mushrooms can require deep fact checking, curating multiple sources of information, and personal experience with the organism.

Of course, generative AIs in the cloud can't have personal experiences with anything. The right way to understand and share knowledge about wild mushrooms is to do it the way the Wurundjeri did. I mentioned embodied knowledge right at the top, and this is an example of that. Sometimes you have to be able to observe something in its context: to smell the air and touch the ground. When reading it in a book isn't enough. Especially when the book was written by a spicy auto-complete.

If we're going to call ourselves information professionals, we need to be thinking a few steps ahead about the consequences of short term convenience. We need to think like a system. We need to think like a spammer. In his conclusion, Finn Brunton thinks about what spam tells us:

Spammers push the properties of information technology to their extremes: the capacity for automation, algorithmic manipulation, and scripting; the leveraging of network effects and vast economies of scale; distributed connectivity and free or very low-cost participation. Indeed, from a certain perverse perspective spam can be presented as the Internet's infrastructure used maximally and most efficiently.

So what is this telling us? Let us imagine.

What happens when generative AI creates a deadly mushroom foraging book and then an automated recommendation algorithm adds it to a standing order for a collections librarian trying to fit two full time roles into one set of working hours? Who is checking whether all of the articles and all of the journals in our big deals even exist, let alone make any sense? If a misconfiguration or a malicious publisher silently deleted all the subject headings or references related to a particular topic, how long would it take your library to even notice?

When you optimise for efficiency, things can be taken so far that there ceases to be any meaning at all.

The people building systems that advise us to add glue to our pizza toppings, eat crushed glass, and drive into the desert aren't interested in context, they're not interested in localised or place-based knowledge, and they're not interested in building human connections.

The information systems I want to work on and with would do more than simply push increasing amounts of text at decreasing amounts of attention.

Among other things, it's this context that makes me worry about the huge firehoses of metadata like Ex Libris's Central Discovery Index or the EBSCO Discovery Service. The efficiency of having a single source of metadata directly from publishers is obviously convenient. But it is in direct contradiction to the push for us to have culturally relevant, decolonised collection metadata.

We can't have it both ways.

The Shock of the Old

David Edgerton is an English historian and educator. His The Shock of the Old is a history of technology in the twentieth century. Edgerton makes the case that the history of technology-in-use provides a much more realistic view of how invention and innovation works, and particularly highlights the importance of repair and maintenance.

The British Library is an interesting case study. Although it doesn't feature in Edgerton's book, we can think with some of his themes to find some lessons for our own libraries.

In October last year, the British Library was hit by a ransomware attack that knocked most of its systems offline for weeks and weeks on end. Some of these systems are still unusable.

When I speak of what Tyson Yunkaporta tells us about allowing interactions to change us, this isn't what I mean!

The Library's official incident review provides details of how the cyber attack came about, but the much more interesting lesson for us is what the consequences were. It reads in part:

The Library's vulnerability to this particular kind of attack has been exacerbated by our reliance on a significant number of ageing legacy applications which are now, in most cases, unable to be restored, due to a combination of factors including technical obsolescence, lack of vendor support, or the inability of the system to operate in a modern secure environment... A few key software systems, including the library management system, cannot be brought back in the form that they existed in before the attack, either because they are no longer supported by the vendor and the software is no longer available, or because they will not function on the Library's new secure infrastructure

• "ageing legacy applications"
• "technical obsolescence"
• "lack of vendor support"
• "no longer supported by the vendor"
• "software is no longer available"
• "will no longer function on the organisation's new infrastructure"

Friends, does any of this sound familiar?

Like all other public libraries in the UK, the British Library clearly has been underfunded for the task it has been given responsibility for. But something I think has been lost in all the commentary about this event is the extent to which this catastrophe was the result of a mismatch between the infrastructure needs of a national library and the commercial interests of software companies.

The average lifespan of a company on the S&P 500 is now only around 20 years. Unless they are tech startups hoping to be acqui-hired, corporations want to do everything they can to make a profit for their shareholders today. They might be gone tomorrow.

Libraries and those of us who work within them have fundamentally different priorities. For all the recent noise about libraries and progressive social values, librarianship is ultimately a very conservative profession. How will we retain our collections and their metadata into the future? Can the data be migrated to a new system, and how long will the current one last? How easy will it be when the inevitable change to preferred terminology occurs, and we want to update our vocabularies? And most fundamentally, how can we guarantee that we always know where everything in our collection is?

It's the job of corporate executives to think about what's needed in the next financial quarter. But librarians often need to think about what's needed in the next generation.

This is why reading the British Library report was both horrifying, and horrifyingly familiar. I could see exactly how they got into this predicament, and I could visualise exactly the systems I'm responsible for that might suffer the same fate.

The ultimate conclusion of the British Library's report was that they are going to protect themselves in the future by moving as many systems as possible to "the cloud". But even they acknowledge that this doesn't really solve the problem, saying:

Moving to the cloud does not remove our cyber-risks, it simply transforms them to a new set of risks.

UniSuper recently learned about this new set of risks, when Google engineers accidentally deleted UniSuper's entire Google Cloud account and all its data.

I predict that Google, Amazon, and even Clarivate will all be long gone before the British Library closes its doors for the last time.

David Edgerton's book helps us to ask questions about the maintainability of systems that have long surpassed their expected lifetimes and the contexts in which they were created.

Will we still be able to access anything after our software as a service vendors declare bankruptcy? What internal processes have we optimised on the assumption that a vendor will take care of it? What happens when they don't? Do we even know how to export and import backups? Does anyone in our team still know how to create a marc record from scratch? Are we training the next generation of librarians who will come after us, and passing on the old knowledge that informs how and why we do things?

We need to ask these questions, because nobody else will.

Technical services and systems work loops and folds back on itself like the Birrarung. Sometimes we're moving forward and backwards at the same time. We use ideas from the nineteenth century or earlier, to build new systems and frameworks for the future. We always have to think about how things from the past will be embedded in the future, and how that affects our present.

So, we need hope.

Hope in the dark

Rebecca Solnit is one of the most skilled essayists of the last century, able to see both the most exquisite details and the sweeping vistas of meaning they are part of. In her collection of essays, Hope in the dark, she writes:

This is an extraordinary time full of vital, transformative movements that could not be foreseen. It's also a nightmarish time. Full engagement requires the ability to perceive both.

I've been talking about some of the frustrating, depressing and exhausting things about librarianship. It is, indeed, both an extraordinary and a nightmarish time. Rebecca Solnit helps us to find a way forward.

One way we can move forward is–as Tyson Yunkaporta teaches us–to reconsider the very idea of moving forward.

The British Library's systems were fragile, and would be difficult or impossible to recover in the case of an attack or other misfortune. The Library's past was always present, in the form of software abandoned by its creators, data formats whose documentation was lost, and security updates that hadn't been performed.

But it's not as if the Library staff didn't know about these things. Quite the opposite: the present they lived in was no doubt full of frustration that they were – for a variety of reasons – unable to remedy these problems. Various futures lived alongside them – from the eventual possibility of opening up software after patents and copyrights expired, to better funding or some kind of artificial intelligence breakthrough. The futures imagined by library staff, by British politicians, and by international hackers interacted with the present of an under-resourced tech team and a couple of basic information security mistakes, with the past of closed-source systems and out-of-business vendors.

If we think in terms of spirals rather than straight lines, if we think of everywhen, then we can more easily understand that pasts, presents and futures exist simultaneously. We can open our imaginations to new possibilities. David Edgerton tells us:

The history of invention is not the history of a necessary future to which we must adapt or die, but rather of failed futures, and of futures firmly fixed in the past. We should feel free to research, develop, innovate, even in areas which are considered out of date by those stuck in passé futuristic ways of thinking.

David Edgerton encourages us to ignore what he calls "passé futuristic ways of thinking". This is a beautiful phrase, but it also captures an approach I think librarians need to embrace more comfortably.

I've lived through a profession-wide panic based on fear of obsolescence. And after a decade or more chasing the latest trends, libraries are now struggling to find people with the kind of deep knowledge about metadata and technical systems that you all know is crucial to running a large library, an d to any claim we have to being a profession. This is not a problem that can be solved through individual hiring decisions, because it's a systemic problem born of a failed future.

We need to fix it together.

I would go further than Edgerton. I strongly encourage you to research, develop, and innovate, especially in areas which are considered out of date by those stuck in passé futuristic ways of thinking.

I've been as guilty as anyone else of venting my frustration by suggesting we should just burn everything down and start again.

But we don't live in a perfect world, and frankly we don't have the resources to start from scratch. We have to make the future whilst operating in the present with the tools we've inherited from the past. As Rebecca Solnit tells us,

Waiting until everything looks feasible is too long to wait.

So what are we going to do? At the beginning of this talk I said I had four things I'd like us to think about:

Lots of things can be true at the same time

Traditional library practice sees like a state, assuming there is "only one thing going on". We need to continue to apply multiple vocabularies to our collections, expressing multiple worldviews. Linked data hasn't yet lived up to its promise, but it might help. However we do it, we'll need to be comfortable with the idea that these different ways of organising concepts and connections between things don't necessarily map neatly onto each other.

Lots of times can be true at the same thing

Everything we do has a past and a future that are active in the present. Maybe multiple futures. Make time to write good documentation for your future self or the person who replaces you. They'll be incredibly grateful. Try to understand why things are set up the way they are, but remember that you don't have to doing things that way. Things that seemed like a good idea at the time ...maybe weren't. And remember that the time that things erroneously seemed like a good idea might be our time, right now. Try to think about the consequences of your decisions in 5, 10, or 25 years. Maybe saving a few dollars this year could end up being very expensive.

Know where you are

Central indexes can be convenient, but we need to be sophisticated in how we use them. We live and work in Australian and New Zealand contexts. Our library collections and their descriptions should reflect that.

Most of our institutions at least pay lip service to the idea that they need to be more culturally aware, and to "decolonise" or "Indigenise". If that's going to happen, we need to be describing our collections in locally contextual, culturally appropriate terms. This isn't something where we can just press a button or tick off a project in a single annual plan. And it's only partially something we can solve collectively with agreed standards. It's a mindset and an ongoing, local responsibility. The people who allocate money and other resources in our organisations don't want to hear this, but that's the reality if we're serious about it.

I don't have any magical solutions to this for you today, but I'd really like for us to keep talking about it and share ideas about what is working in your context.

Share what you know

ANZREG is a wonderful community that shares freely. But it seems to me that we could be sharing more ideas, tools, and techniques in librarianship generally. I encourage you all to be a little more brave. Send that "dumb question" to the email list. Publish that blog post you're not sure about. Post some code to GitHub. Say yes when someone invites you to give a conference talk. You're good enough to write a journal article. You know enough to peer review a conference talk or a paper.

Tyson Yunkaporta quotes his friend Katherine Collins in his latest book Right Story, Wrong Story. She says:

When learning new things, we are trained to think Is this true or false? But it is so much better to think When will this be useful? Also When should I not rely on this? When will it fall apart?

These are good questions to think about as we attend the sessions this week. Let's make connections, and let them change us, and think about how in turn we're going to change the systems we work in and with and on.