iiinova — where your digital self keeps evolving

Xing Shi Cai @docilelargeness.bsky.social · 27/07/2026

I read this and asked a machine (AI/LLM) "What are we going to do, if machines can think better than us?" It suggested me to pick up a hobby like pottery, blacksmith, or cooking. It even give me a map of places where I can learn blacksmith near my IP address. substack.com/home/post/p-...

substack.com

The Dark Night of Mathematics

What are we really doing?

010

Xing Shi Cai @docilelargeness.bsky.social · 26/07/2026

I was very happy when Anthropic handed out $100 gift credits. Then one overnight session --- fixing a bug in my VPN setup --- cost ~$82. Moreover, because the session broke my VPN, I had to install a second VPN in order to use Fable to fix the first VPN it broke.

000

Xing Shi Cai @docilelargeness.bsky.social · 27/03/2025

Asking it to make the picture "much more detailed and awe inspiring" gives the following result. I don't feel the magical feeling I had the first time I saw AI generated pictures. But the progress is still quite impressive.

020

Xing Shi Cai @docilelargeness.bsky.social · 27/03/2025

The first one was generated last fall (2024) via Microsoft Copilot (DALL-E 3.) I asked the latest GPT-4o to make it more photo-realistic and got the second picture, but it missed the helmet. Asking it to add the helmet turned the bunny's head the wrong way. One more prompt fixed this.

110

Xing Shi Cai @docilelargeness.bsky.social · 27/03/2025

I asked #llm gpt-o3-mini-high, claude-3-7-sonnet, grok3, gemini-2.5-pro-exp and deepseek-r1 to "Create latex code using tikz package to draw a cute bunny emoji".

110

Reposted by Xing Shi Cai

Peter Singer @petersinger.info · 06/03/2025

The latest episode of “Lives Well Lived”, the podcast I co-host with Kasia de Lazari-Radek, is now available. In this episode, we talk to Robert Sapolsky, the renowned neuroscientist, primatologist, and author who has spent decades studying baboons to understand how stress affects health. (1/4)

youtube.com

Robert Sapolsky on Free Will

YouTube video by Peter Singer

174

Reposted by Xing Shi Cai

CEMUS Uppsala University @cemusuppsala.bsky.social · 11/03/2025

Save the date Nov 20: ClimateExistence 2025 Symposium in Uppsala We want to welcome you all to a special edition of ClimateExistence, this time as a half-day symposium and evening social in Uppsala, so save the date November 20 after lunch! Read more: www.cemus.uu.se/nov-20-ce-20...

011

Reposted by Xing Shi Cai

Kevin Anderson @kevinclimate.bsky.social · 11/02/2025

Giving an open lecture at Uppsala University tomorrow morning (10.15-12.00CET - Wed' 12th Feb); it will also be online at uu-se.zoom.us/j/65559664314 The event is part of a @cemusuppsala.bsky.social lecture series. For more details see: www.cemus.uu.se/feb-12-ander...

uu-se.zoom.us

Join our Cloud HD Video Meeting

Zoom is the leader in modern enterprise video communications, with an easy, reliable cloud platform for video and audio conferencing, chat, and webinars across mobile, desktop, and room systems. Zoom ...

0128

Xing Shi Cai @docilelargeness.bsky.social · 09/02/2025

I do have a workstation which can sort of run Llama 3.3. But for my laptop that's impossible. In any case, if I do want to play with them, using Groq for free now.

000

Xing Shi Cai @docilelargeness.bsky.social · 08/02/2025

You mean The Lifecycle of Software Objects? It's a science-fiction novella, and a diverting read, to be sure. Though, I recall a review, I think from a PhD student, who said something like, "Reading this book reminds me too much of my research. So, not as relaxing."

020

Xing Shi Cai @docilelargeness.bsky.social · 08/02/2025

I see. Sorry that I misunderstood. I wish we had an LLM inside Bsky to check posts before we submit them. 😁

120

Xing Shi Cai @docilelargeness.bsky.social · 08/02/2025

This makes me think—explaining things in class might not be as crucial for teachers anymore. Maybe our real role is to make things more engaging for students instead. That’s why I’m spending less time in class explaining details about computations, and more time telling stories about math.

010

Xing Shi Cai @docilelargeness.bsky.social · 08/02/2025

Since these models started getting better at math last year, the number of math questions I get from students has dropped to almost zero. Either I’ve gotten way scarier, or they’ve found better ways to get help.

020

Xing Shi Cai @docilelargeness.bsky.social · 08/02/2025

o1 was the first model that could solve math problems that top students find easy, but were tricky for everyone else. Then came QwQ, Gemini 2.0 Flash Thinking, DeepSeek-R1, and now O3. It's wild to that it is now very hard to find a math problem students can do that AI can’t.

210

Xing Shi Cai @docilelargeness.bsky.social · 08/02/2025

AI studio is really not a good product name these days. Too many similar names out there, like AI Studios, AI Studio, AILab Studio, and Studio AI.

100

Xing Shi Cai @docilelargeness.bsky.social · 08/02/2025

I did pledge to not to login to bsky again tonight 15 minutes ago. 😩

000

Xing Shi Cai @docilelargeness.bsky.social · 08/02/2025

Have you read The Lifecycle of Software Objects by Ted Chiang? It's about humans raising AI pets but eventually getting bored and freezing them on a hard drive. It’s about a similar ethical issue. Though here, humans can also stop the toys from feeling upset if they want — or choose not to.

030

Xing Shi Cai @docilelargeness.bsky.social · 08/02/2025

I haven't found a reason to use any local models. Every time I try, I keep asking myself why I do use the commercial ones which are better and faster? Moreover, I have already paid for them with a subscription fee.

100

Reposted by Xing Shi Cai

Dr. Aaron Thierry @thierryaaron.bsky.social · 07/02/2025

So which is it? "Puzzles", "Mystfies", "Stuns", "Shocks"... Makes quite a difference, no? I'd opt for "Radicalises" 😉

74011

Reposted by Xing Shi Cai

Peter Singer @petersinger.info · 07/02/2025

There is an old saying, "Let justice be done, though the heavens fall!" It makes people feel good about themselves that they are taking the moral high ground, and they can get outraged at everyone else. (1/5)

67163

Reposted by Xing Shi Cai

Peter Singer @petersinger.info · 09/01/2025

Competition Time: Celebrate Season 2 of Lives Well Lived! This week, we kicked off Season 2 of my podcast, “Lives Well Lived” which I co-host with Kasia de Lazari-Radek. To celebrate, I invite you to share an action you’ve taken to improve the world. (1/3)

2153

Xing Shi Cai @docilelargeness.bsky.social · 15/01/2025

Keep going. But don't be attached to the outcome. The universe will run it's course and we have to accept that. This is what I am trying to do.

030

Xing Shi Cai @docilelargeness.bsky.social · 15/01/2025

If we want social media to belong to the people, we really should all go to Mastodon. We do not even need that many people. Just all the environmental activists of the world would form a very interesting online community.

000

Reposted by Xing Shi Cai

George Monbiot @georgemonbiot.bsky.social · 14/01/2025

On CBC last night, I spelt out the moral case to keep on trying, however hard it might be. Defending the living planet is not an option we can abandon. Ultimately, it is all that counts. www.youtube.com/watch?v=qX2V...

youtube.com

Author George Monbiot on countering despair amid alarming climate change trends

YouTube video by CBC British Columbia

551415472

Reposted by Xing Shi Cai

Dr. Aaron Thierry @thierryaaron.bsky.social · 22/12/2024

In a world where the dark, dystopian visions of Cyberpunk have long dominated our sci-fi imaginations, a new genre emerges as a beacon of hope: Solarpunk youtu.be/UVlBmdvIC6s

youtu.be

SolarPunk Cities: Our Last Hope?

YouTube video by DamiLee

16118

Xing Shi Cai @docilelargeness.bsky.social · 21/12/2024

Have you tried it yourself? Even gpt-4o can do this. LLM are bad at many tasks. But this sort of things are getting really easy.

To find out how many kiwis Oliver has, let's break down the problem step by step:

- On Friday, Oliver picks **44** kiwis.
- On Saturday, he picks **58** kiwis.
- On Sunday, he picks double the number of kiwis he picked on Friday, so that's **2 × 44 = 88** kiwis.

However, we are told that **five of them were a bit smaller than average**, but this doesn't change the total count of kiwis. So, he still picks **88** kiwis on Sunday.

Now, let's add up the total:

- Friday: **44**
- Saturday: **58**
- Sunday: **88**

Total = 44 + 58 + 88 = **190** kiwis.

So, Oliver has **190** kiwis in total.

100

Reposted by Xing Shi Cai

Plant Futures DKU @plantfuturesdku.bsky.social · 08/12/2024

Starting a new tradition: sharing weekend. This week: Elizabeth Kolbert's When the Arctic Melts, exploring Greenland’s ice cores, rapid climate shifts, and the lives of scientists studying them. Read: www.newyorker.com/magazine/202... #ClimateChange #Science #Arctic #WeekendReads

newyorker.com

When the Arctic Melts

What the fate of Greenland means for the rest of the Earth.

021

Reposted by Xing Shi Cai

Guardian NUJ @gonuj.org · 04/12/2024

Today Guardian and Observer journalists are on strike over plans to sell the Observer to a loss-making digital start up and put its journalism behind a paywall. 🧵 How can you support us? SHARE our posts WRITE to guardian.readers@theguardian.com JOIN US - there's cake!

7425971452

Reposted by Xing Shi Cai

Peter Singer @petersinger.info · 04/12/2024

Ingrid Newkirk, the co-founder and president of People for the Ethical Treatment of Animals (@officialpeta.bsky.social), talks about the disturbing reality of animal experimentation, highlighting the many animals subjected to tests that cause them to suffer without benefitting humans.

youtu.be

Ingrid Newkirk on Animal Experimentation

YouTube video by Peter Singer

1155

Reposted by Xing Shi Cai

Ethan Mollick @emollick.bsky.social · 03/12/2024

This used to be a very hard problem, I have seen a number of startups try (and fail) to solve it over the years. Now it is suddenly pretty trivial with open tools (though I am not sure why Maria got a new haircut)

6423

Xing Shi Cai @docilelargeness.bsky.social · 03/12/2024

This is a bit misleading. Mastodon is a much closer to email. It does look like there will be another bluesky protocol provider any time soon. It could very very easily become another Twitter.

010

Xing Shi Cai @docilelargeness.bsky.social · 03/12/2024

This is the first ever post to use #dku and #dukekunshanuniversity on bluesky.

000

Reposted by Xing Shi Cai

Plant Futures DKU @plantfuturesdku.bsky.social · 03/12/2024

Plant Futures #DKU Spring Recruitment has started! Check our website: tinyurl.com/pfdku to know more about us.

001

Reposted by Xing Shi Cai

Simon Willison @simonwillison.net · 03/12/2024

Ultimately the thing I care about most is if a human has evaluated the message and found it to be high enough quality that they're willing to stake their own credibility on what it says Provided the message accurately represents them I'm not too bothered about how they produced the individual words

7685

Xing Shi Cai @docilelargeness.bsky.social · 03/12/2024

Messages focusing on ethics and environmental issues have had very little impact. So I think perhaps focusing on the consumerism aspect of vegetarianism would work better.

000

Xing Shi Cai @docilelargeness.bsky.social · 03/12/2024

I think it's more important that what has been written has some benefits to the readers. Who wrote them doesn't really matter.

000

Xing Shi Cai @docilelargeness.bsky.social · 01/12/2024

When I teach derangement in combinatorics I usually mention what it means in everyday English and use this book as an example, hoping some of them might get curious.

020

Reposted by Xing Shi Cai

Dr. Aaron Thierry @thierryaaron.bsky.social · 01/12/2024

“Let us make no mistake: the climate crisis is also a crisis of culture, and thus of the imagination" - @amitav.bsky.social

15716

Reposted by Xing Shi Cai

𝕁𝕦𝕝𝕚𝕒𝕟 ℝ𝕚𝕕𝕕𝕖𝕟 @eduridden.bsky.social · 27/11/2024

Interesting #OpEd in today’s #SMH talking the impact of #GenAI on #higherEducation that I found triggering. The issues it highlights is the issue that how we assess is broken. The degree is not worthless, our assessment methods are. Time to change the system. #EduskyAi #AIEdu

smh.com.au

Cheating is now so rampant that uni degrees have become worthless

The university as we know it is dead. COVID-19 nearly killed it, and AI has dealt the final blow.

083

Xing Shi Cai @docilelargeness.bsky.social · 01/12/2024

I wish Claude allows publicly sharing a conversion with a link, just as ChatGPT and Gemini does. I have to say, given how much resources these companies have, the improvement of UI has been slow.

030

Reposted by Xing Shi Cai

Ethan Mollick @emollick.bsky.social · 30/11/2024

“Claude, we are mice (pretend). You are a mouse mastermind. Come up with a way to reach the flower without getting caught“ Solid plan. Then I add a complication.

7633