"Just look at that 3-year TOC. Campfire running on Hetzner would cost you $479 for 500 users over three years ($299 + 36 * $5). Slack with the same number of seats would cost you $270,000!!! The ONCE solution is literally 99.9% cheaper. It’s crazy." world.hey.com/dhh/campfire-i…

Our brief thoughts on Large Language Models in Education, and what we’re doing about it 👇 unsw.edu.au/news/2024/04/t…

The Orb by MARC FORNES / THEVERYMANY from my recent Seminar visit to the Google campus in California, made up of 217,847 rivets assembled by hand.

Perfect day to be speaking at UNSW Computing!

Good to talk all things AI and its role in the classroom to mark the Day of AI today on ABC Sydney

Jake Renzella ABC Sydney #DayOfAI

If you're teaching NLP, do you start with Transformers from the outset, or focus on foundations and "get to the cool stuff" later? Aditya Joshi + team's new preprint that explores this, and more: arxiv.org/html/2405.0985…

The #DeadInternetTheory’ makes eerie claims about an #AI-run web. The #truth is more sinister theconversation.com/the-dead-inter… #fintech #ArtificialIntelligence #MachineLearning #GenerativeAI #GenAI Jake Renzella UNSW @UniMelb The Conversation - Australia + New Zealand

The ‘dead internet theory’ makes eerie claims about an AI-run web. The truth is more sinister theconversation.com/the-dead-inter… via The Conversation - Australia + New Zealand

Is most of the content on the internet fake? Jake Renzella (UNSW) and Vlada Rozova (@UniMelb) explain what the dead internet theory means – and why we should be warier about how we're manipulated for profit and political gain. theconversation.com/the-dead-inter…

Is most of the content on the internet fake? Jake Renzella (UNSW) and Vlada Rozova (@UniMelb) explain what the dead internet theory means – and why we should be warier about how we're manipulated for profit and political gain. theconversation.com/the-dead-inter…

Surreal to see my The Conversation - Australia + New Zealand article on the Dead Internet Theory make it to the front page of Hacker News: news.ycombinator.com/item?id=406432…

I get wicked car sickness, might be the first time I update to an iOS beta fairly early

We’ve seen in our research (dl.acm.org/doi/abs/10.114…) the propensity for models to ignore prompt instructions and provide complete solutions to students. Despite benefits, this breaches what we know about how learning occurs

We design a new voting rule called the Solid Coalition Refinement (SCR) rule that simultaneously satisfies committee monotonicity and Dummett’s Proportionality for Solid Coalitions property. cse.unsw.edu.au/~haziz/SCR.pdf

Im #hiring! Looking for a postdoc Research Fellow in optimisation and RL with a strong publication record on reinforcement learning, ML, dynamic scheduling, optimisation. Ideally, this candidate would also have an experience in building prototype demo/systems and/or sandbox for