#digitalsocialscience

2025-10-13

We're hiring! Three pivotal new roles tailored to help our new HASS Digital Research Hub build capability in HASS digital research at the Australian National University, and nationally in partnership with the Australian Research Data Commons (ARDC):

Research Fellow (Computational Methods)
jobs.anu.edu.au/jobs/research-

Research Fellow (Values Based Digital HASS)
jobs.anu.edu.au/jobs/research-

ARDC RSE-CEP Research Software Engineer
jobs.anu.edu.au/jobs/research-

#digitalhumanities #digitalsocialscience #rse

2025-05-19

It's nice to be back in California. Today I'll be talking at D-Lab Berkeley about what I've learned after a year in Australia!

'Antipodal Experiments: Digital Humanities & Social Science in Australasia': dlab.berkeley.edu/events/antip

#d-hass #digitalhumanities #digitalsocialscience #rse

2025-04-08

The ANU HASS Digital Research Hub is advertising a research software engineering position, to contribute to the Social Science Research Infrastructure Network (SSRIN) project, funded by ARDC and led by University of Queensland. It would be a great role for someone interested in computational social science and/or research software engineering.

jobs.anu.edu.au/jobs/research-

Please circulate to your networks.

#socialscience #digitalsocialscience #computationalsocialscience

2025-02-06

Help me build a research software engineering (RSE) team dedicated to the humanities and social sciences! Two key roles based in Canberra, Australia, for #softwareengineers keen to make a difference:

Principal Research Software Engineer (RSE): jobs.anu.edu.au/jobs/principal

Senior Research Software Engineer (RSE): jobs.anu.edu.au/jobs/senior-re

#digitalhumanities #digitalsocialscience #rse

2025-01-10

The misuse of the concept of assemblage within digital social science

I’ve seen a growing trend to use the Deleuzian concept of assemblage in a way that fails to distinguish between internal and external relations. What distinguishes an assemblage from an entity or a totality is these elements are externally related, rather than the internal relation between elements which jointly constituted a whole. Or at last they are a mix of internal and external relations. But what an assemblage is not is a purely internally related entity. That’s the whole reason for coining the neologism in the first place, to distinguish it from a being/whole. It’s the difference between a relation which cannot exist apart from through its connection to a whole, as opposed to a relatively autonomous relation that can manifest in a different form. So for example the battery on my iphone has an internal relation to the phone, whereas the airpods have an external relation to it. You could extract the battery and install it somewhere else but it would have to enter a constitutive relation with another iphone, whereas the airpods can be used in relation to a range of entities.

This matters because if we treat assemblages as containing internal relations, we are basically talking about totalities while imagining we are doing the opposite. The analytical virtue of assemblage theory is that it helps us sketch out how heterogeneous elements are drawn together into complex coalitions of existing things, able to form and reform in dynamic and multifaceted ways. If you treat the assemblages as if they have internal relations then you are suddenly imagining vast totalities which loom across the social world, without recognising the dynamic character which is the whole analytical point of the theory. In this sense if you imagine ‘AI’ as an assemblage, without making this distinction, it becomes this vast and impenetrable juggernaut which rampages obscenely across a world which it transforms. It becomes a mega force, to use Filip Vostal’s phrase, to which we must either subordinate ourselves or reject in its entirety.

If you treat an assemblage as an internally related thing you’re effectively just connecting up a load of heterogeneous elements and saying ‘this is a thing’. But things being connected don’t make them a thing, even in terms of assemblage theory. What matters is how they’re connected, under what circumstances, with what results. This is why assemblage theory is powerful but you completely miss that if you lose the internal/external distinction, whether explicitly in terms of conceptualisation or tacitly in terms of how you’re narrating the analysis.

I’ve wondered recently why some critics of generative AI see it as uniquely obscene, as opposed to another expression of platform capitalism. Fundamentally I don’t see it as any different on a moral level (clearly technologically and potentially socially it is very different) to watching YouTube, posting on Instagram or using Uber. Or for that matter driving or taking flights. It involves a complicity in a system you would rather was different, licensed by the affordances of that system. You can respond to this in multiple ways but these responses tend to be connected into the practical logic of living a life e.g. non-drivers sometimes using Uber. I often agree with much of the substance of the critique, yet don’t end up in the same place politically and morally as they do. I wonder if there might, at last sometimes, be a disagreement about the ontology of ‘AI’ underpinning this. To the extent its seen as a totality, whether deliberately or through its misattribution as an assemblage, I part ways at an analytical level and I think a difference of moral and political opinion follows from this.

#AI #assemblage #deleuze #DigitalSocialScience #generativeAI #LLM #socialOntology

2024-06-11

Our AI as Infrastructure (AIINFRA) project has a website: aiinfra.anu.edu.au/.

AIINFRA is an experimental project. The goal is to design and build a prototype open-source LLM tool tailored for transnational (Australia, Aotearoa NZ, UK) historical research, but this will be in service of the primary goal of understanding the technical potential of LLMs and developing test categories appropriate to the academic and GLAM communities.

#AiResearch #digitalhumanities #digitalsocialscience

2024-03-05

This is a key role within our AI as Infrastructure (AIINFRA) project, exploring the use of LLMs for transnational historical research (Australia, Aotearoa / New Zealand, United Kingdom).

Please forward to likely candidates!

jobs.anu.edu.au/jobs/senior-re

#digitalhumanities #digitalsocialscience #digitalhistory

2024-02-26

It's great to see @kingsdh are advertising for a Lecturer / Senior Lecturer in Digital Products & Industries to complement their 'Career Accelerator' in the subject.

I'm biased - I was involved in development of the career accelerator and offered a course there in DH product development - but am convinced it's an important area for future digital HASS. We need to be in the engine room of digital design.

kcl.ac.uk/jobs/085006-lecturer

kcl.ac.uk/short-courses/kings-

#digitalhumanities #digitalsocialscience

2024-01-28

It's also worth noting that we get more info from the command line and of course the code itself with langchain (and presumably other frameworks), which amounts to documentation of the tool's state at the time of the query.

It suggests there is some hope for the definition and implementation of transparency standards (research software engineering / technical and scholarly).

#digitalhumanities #digitalsocialscience #digitalhistory
#langchain
#rse

Screenshot of the gpt-researcher LLM tool's (default) configuration.
2024-01-28

I'm cautiously optimistic about transparency tools we can use to build bespoke LLM products for the humanities & social sciences.

These screenshots show what the langchain.com/langsmith tool exposes behind the scenes of a simple historical query to the github.com/assafelovic/gpt-res research tool.

gpt-researcher is running locally on my machine, using OpenAI and Tavily services, but uses open source code so this could be changed.

#digitalhumanities #digitalsocialscience #digitalhistory
#langchain

Output of the 'gpt-researcher' AI tool, for the query 'The history of Australia federation, leading up to 1901'. The research report includes articles from WIkipedia and the National Library of Australia, describing their relevance, reliability, and significance.A screen shot of the langsmith Large Language Model (LLM) monitoring tool, showing the output of the system behind the scenes.A screen shot of the langsmith Large Language Model (LLM) monitoring tool, showing the output of the system behind the scenes.A screen shot of the langsmith Large Language Model (LLM) monitoring tool, showing the output of the system behind the scenes.
2024-01-17

It's wonderful to have started my new role at ANU! I'll be taking it slowly but am excited about helping enable digital HASS research locally, nationally, and regionally. Lots to learn, and lots of people to meet 😀.

#rseng #infrastructure #digitalhumanities #digitalsocialscience

Client Info

Server: https://mastodon.social
Version: 2025.07
Repository: https://github.com/cyevgeniy/lmst