Antony Tibbs
antonyt.bsky.social
Antony Tibbs
@antonyt.bsky.social
Product Owner - AI @ ECU. // AI, education, technology, innovation, productivity, scifi // Views my own. // I write a newsletter called Tachyon: https://tachyon.substack.com
No need to sign in to Miro, just follow the link and pan around to read the information. Each development includes some info on how to access, an activity, and links to read more.
2/
February 12, 2025 at 4:24 AM
The classification of 'augmentation' vs 'automation' is interesting, as is the fact that augmentation is more common. But the data excludes API usage, which I would expect to include a lot more automation-type tasks.
February 10, 2025 at 10:51 PM
For people working in education the top task (but not by much of a margin) was "Design and develop comprehensive curricula and materials". No surprises there.
/3
February 10, 2025 at 10:51 PM
It shows the kinds of research that can be done when you not only have access to this kind of data but also the ability to build a research pipeline that leverages genAI to process and categorise data at scale.
/2
February 10, 2025 at 10:51 PM
Anthropic has just released the Anthropic Economic Index, which is really all about the tasks that people perform with Claude. Computer science by far the most common job type.
/1
#AI #productivity #education
February 10, 2025 at 10:51 PM
A sneak peak at the (almost finished) Miro board for my workshop tomorrow at #WATLF2025. This will be the main activity - an interactive timeline of major developments in generative #AI from the release of ChatGPT in Nov 2022 to now. Kind of hard to believe all of this is just from the past 2 years.
February 5, 2025 at 8:08 AM
The sport section label in my (formerly) local second hand bookshop in Melbourne. Sick burn from the owner, I mean if you must have a sports section in your shop…
December 6, 2024 at 11:09 PM
In case you missed it, here is our poster: "Empowerment and Connection: Developing organisational capacity in AI literacy through connectivist pedagogy".
publications.ascilite.org/index.php/AP...
#AI #education #HE #ascilite24
December 4, 2024 at 5:36 AM
“It’s about the teaching not the tech” - Chris Honig. A universal truth #ascilite2024
December 4, 2024 at 3:21 AM
I like this statement of beliefs and strategy from Chris Honig at the start of his presentation. Agree - we need lots of experimentation #AI #ascilite2024
December 4, 2024 at 3:05 AM
Michelle Pedlow and Justine Maldon (ECU) share results from an intervention based on TEL and transition pedagogies to enhance student digital literacies. Some interesting data on student confidence in #AI use for learning and assessment too
#ascilite2024 publications.ascilite.org/index.php/AP...
December 4, 2024 at 2:49 AM
Kate Tregloan (UniMelb) - a group/teamwork lens on assessment design allows for clearly defined roles for students and AI from individual to collaborative tasks. Including #AI use in rubric incentivises students to disclose how they use it

#ascilite2024
publications.ascilite.org/index.php/AP...
December 4, 2024 at 1:06 AM
Interesting to see how #AI (using Cogniti) can be integrated into teaching workflows to provide triage and early warning of group work at scale - smart work from Angela Sun and Helen McGuire (USyd) #ascilite2024 publications.ascilite.org/index.php/AP...
December 4, 2024 at 12:38 AM
Great use of structured prompting from @d4nl.bsky.social and Helen Patterson in the creation of their branching nursing simulation in Cogniti. Used 'expected responses' to help ground the agent #ascilite2024 #AI doi.org/10.14742/apu...
December 4, 2024 at 12:35 AM
@jasonmlodge.bsky.social Education fit for a vulcan but not a human? What’s at stake for ed tech research is avoiding a tech-maximalist future where humans no longer learn from humans

#ascilite2024
December 3, 2024 at 11:00 PM
I forgot to share my pics #ascilite2024
December 3, 2024 at 3:11 AM
Fernando Jativa presenting on an awesome #VR 3D annotation app. Loved the user-centred design approach they took to develop the tool #ascilite2024
December 3, 2024 at 12:37 AM
TEQSA just released their #GenAI for #HE toolkit that summarises best practices from the Request for Information. Responses from 202 Australian institutions, organised into three sections: Process, People and Practice. www.teqsa.gov.au/guides-resou...
November 28, 2024 at 4:02 AM
According to Cloudflare, the vast majority of their customers opt to block the crawlers that #AI companies use to scrape training data. blog.cloudflare.com/...
October 1, 2024 at 11:07 PM
What accuracy rate should we accept when relying on #AI to summarise information as part of our workflows? 100%? 90%? What level of accuracy would you expect and accept from a human?
September 30, 2024 at 12:00 AM
I conducted an informal experiment into the accuracy and quality of #AI summaries. I found that they often contain mistakes, model matters for accuracy and overall quality, and that I nonetheless prefer Claude 3.5's summaries more often than my own. Read more: tachyon.substack.com...
September 28, 2024 at 10:00 AM
I made my first 'mini software' utility using Claude Artifacts. Needed to compare options for room names with a 33 character limit. Plug in the different options and get instant character counts - much more satisfying than messing around in Word! Try it: claude.site/artifacts/ef...
#AI #productivity
September 11, 2024 at 12:19 AM
My next Tachyon post will be a recap of how people have been making sense of gen AI over the past year or so. Here's a behind-the-scenes peek of a timeline of the shortlisted pieces.

You can subscribe now to read the post in full when it drops: tachyon.substack.com
December 29, 2023 at 6:29 AM
These are the top three tags in my Readwise Reader library at the end of 2023. I read 335 of the items saved in the first tag and 263 from the second. I'll be sharing the most interesting reads in my next Tachyon post.
December 28, 2023 at 11:18 AM
Enjoyed Ethan Mollick’s latest post summing up the current state of play with regards to LLMs www.oneusefulthing.org/p/the-shape-...
October 4, 2023 at 9:42 AM