Twitter Beğeni Hilesi

On Thu 19 Sept, 2-3pm (BST), we have Dr MartinCompton joining us to share his 30+ years in education & how his caring heart has led to his passion in equity-centred education! More info: bit.ly/MCompton0924 #HEresearchers #TELresearchers

On Tue 15 Oct, 9:30-10:30am (BST) / 7:30-8:30pm (AEDT), we'll visit Prof. Phillip Dawson in Australia to learn more about the challenge of being both ethical and strategic in a research career! More info: bit.ly/PDawson1024 #HEresearchers #TELresearchers

Amongst a lot of barely coherent an ill-informed twaddle (not least in education press) it’s good to see an article in national media that is coherent and rational about generative AI. Acceptance is key to adjustment theguardian.com/commentisfree/…

85 years. Messes with my brain

Last week marked the end of the Afe Babalola Centre's two-week residential at King’s with 30 teachers from across Africa. Teachers visited King’s and TEDI-London to co-develop a training program which will upskill educators and teachers at scale. ⬇️ kcl.ac.uk/news/afe-babal…

New #WeeklyResourceRoundup of events and resources useful for those teaching in HE just published, some great toolkit and online events there. wordpress.aber.ac.uk/e-learning/202… #AberPGCTHE

God, this report 😔 #AccentBias #Academia

Most of the work in AI that has any real impact shows up in pre-prints or on blog posts. By the time, it appears in a journal with an "Impact Factor", it is obsolete. What does it tell us about academic publishing?

Specialist Cordelia McGee-Tubb has said accessibility is like blueberry muffins: You can’t bake muffins and then try to add blueberries after the muffins are baked. Similarly, you can't have accessible websites if you try to add accessibility at the end. It must be baked in.

What I don’t get is. If you’re an academic, how is it possible to be sent reasonable adjustment plans and then refuse to implement them or say out loud “oh I’ve not read that” and not lose your job?

🇩🇪In Berlin for a conference on #LLMs and philosophy and just received a desk reject from a journal that said "we have yet to be persuaded that there is anything of philosophical significance with this technology." Last time this happened was in 1994 for an essay on the Internet.

And still half the folk talking about it say ChatGTP.

So interesting and also the thread that follows. Shines a light on what we value and how that privileges some over others. Not AI related …. Yet it is in a way

Lincoln Michel I would have liked more nuance. What artist doesn’t use Auto-Tune? What film isn’t post processed using AI/ML? What book hasn’t had NLP editing tools? Take the generative label off of it and “AI” is a fundamental, even ubiquitous tool for nearly all artist.

So there is a paper from the Australian government on how AI isn’t good at summarization (aph.gov.au/DocumentStore.…). But that doesn’t match other research, like the content below. It turns out the Australia paper used old Llama 2-70B & Mistral 7B. Using a frontier model matters.

I’ve realised a very useful AI enhancement to the kindle (&audible come to think of it) would be a summary up to the point you’ve got to- particularly when you restart a book after a while

If society would have just accepted what people were saying in their actual voices instead of telling them it was Bad English, we might not have this problem in the first place.

Lean into the cognitive dissonance! Something can be bad and good simultaneously. When bubbles burst we still have the liquid to make more

Hey King's College London, whether you've been here for a while or are about to join us. Don't buy any books. Check your library first. We've been building your/our collection since 1829, chances are *very high* we already have it. 👉 librarysearch.kcl.ac.uk

Bad news imo that Recast is folding. So rare that I subscribe with money to something.. is the bubble bursting??