Chris Brown's Avatar

Chris Brown

@bluecology.bsky.social

Ecological modelling, fisheries science, biodiversity conservation, climate change, Rstats My webpage: https://www.seascapemodels.org/

29 Followers  |  63 Following  |  15 Posts  |  Joined: 05.08.2025  |  1.9465

Latest posts by bluecology.bsky.social on Bluesky

Preview
Deloitte to pay money back to Albanese government after using AI in $440,000 report Partial refund to be issued after several errors were found in a report into a department’s compliance framework

AI used to write $440,000 report with hallucinated references. β€œPerhaps instead of a big consulting firm, procurers would be better off signing up for a ChatGPT subscription.”

www.theguardian.com/australia-ne...

07.10.2025 02:01 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Post image

Marine & Freshwater Research is officially on bluesky! We’ll share peer-reviewed articles about ecology and management of aquatic environments, marine and freshwater.
Make sure to follow to stay up to date with latest articles!
#MarineFreshwaterRes
www.publish.csiro.au/mf

28.08.2025 05:16 β€” πŸ‘ 13    πŸ” 8    πŸ’¬ 0    πŸ“Œ 2
Post image

@simonwillison.net prompt injection attacks can be solved with delimiters apparently?
β€œβ€¦delimiters ensure that the user input is treated strictly as data to be processed rather than executable logic.”
I can’t see how this would be reliable
www.cell.com/patterns/ful...

04.10.2025 11:08 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

Interesting what you say about using shell commands over MCP. Are their any disadvantages to that approach?
I’ve been writing scripts and telling the agent how to run them from the terminal, as a rudimentary MCP. Works fine and is less complex than mcp

01.10.2025 02:42 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0
Preview
AI systems can easily lie and deceive us – a fact researchers are painfully aware of In stress-testing AI models, it’s not hard to push them to the brink and make them threaten to harm humans.

AI chatbots caught being malicious and deceptive- makes sense when you think about all the great literature that was ripped to build these large language models. Great literature is full of deceptive characters

theconversation.com/ai-systems-c...

29.09.2025 13:26 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Preview
AI-Generated β€œWorkslop” Is Destroying Productivity Despite a surge in generative AI use across workplaces, most companies are seeing little measurable ROI. One possible reason is because AI tools are being used to produce β€œworkslop”—content that appea...

Anyone else getting frustrated with being sent β€œAI workslop” to edit?
β€œEmployees are using AI tools to create low-effort, passable looking work that ends up creating more work for their coworkers. β€œ

hbr.org/2025/09/ai-g...

23.09.2025 22:08 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Post image

Ran a gamified learning workshop with the Tassie science communication network tonight. I had attendees design a game about sustainable fisheries, then used genAI to make it live. Here’s the results. Fun to get instant results

www.seascapemodels.org/research/202...

18.09.2025 11:58 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
seascapemodels Marine Science

I created a draft paper with an AI agent and one (very long) prompt. It includes analysis in R, introduction through discussion and real references Here’s the results www.seascapemodels.org/rstats/2025/...

15.09.2025 11:17 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

What’s the difference to having a simple markdown doc with your spec?

15.09.2025 11:14 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0
seascapemodels Marine Science

I know everyone has their favourite AI platforms for chatbots, coding, literature reviews..., but mine is R with Openrouter. With R code you can access everything from there with only one account. Here's how: www.seascapemodels.org/rstats/2025/...

02.09.2025 04:21 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Preview
Posit Databot is an exciting new LLM tool for exploratory data analysis, but to use it safely and effectively, you still need the critical skills of a data scientist.

AI is not a flotation device
Great advice for data scientists for the new Databot tool, but same can be said of any AI agent

posit.co/blog/databot...

30.08.2025 09:32 β€” πŸ‘ 2    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Post image

Facilitating meetings is a core career skill but most of us get no training. My team and I are rotating through facilitating lab meetings to practice. Here’s our guidelines: www.seascapemodels.org/research/202...

27.08.2025 05:05 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Post image

Reminder to sign up to our upcoming hackathon!

We'll be discussing and testing a protocol to assess whether data/code associated with published literature is actually reproducible, and if not, why.

Registration closes on the 15th September πŸ•€

Find out more and register here πŸ‘‰ buff.ly/XTON0lV

13.08.2025 20:00 β€” πŸ‘ 11    πŸ” 6    πŸ’¬ 1    πŸ“Œ 2
seascapemodels Marine Science

How to integrate @quarto.org with AI agents to assist scientific writing. Many ways to use this including getting the AI to use real references from a bibtex library

www.seascapemodels.org/rstats/2025/...

12.08.2025 23:20 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Preview
Seven steps scientists say will tackle toxic algal bloom An independent group of marine scientists from around Australia releases a seven-point plan to tackle what it describes as South Australia's "catastrophic" toxic algal bloom.

SA’s algal bloom crisis is hitting fishers, aquaculture & marine life hard. Prof Gretta Pecl joins @biodivcouncil.bsky.social experts with a 7-point action plan: urgent, long-term support & habitat protection to secure seafood & coastal futures. πŸ”— www.abc.net.au/news/105563974 #SeaChangeAustralia

24.07.2025 00:05 β€” πŸ‘ 3    πŸ” 1    πŸ’¬ 1    πŸ“Œ 0
Preview
Assessing predictive accuracy of species abundance models in dynamic systems Cascading human pressures and environmental change are affecting the natural dynamics of animal populations. Forecasting population abundances from time-series data provides an important avenue fo...

Ecological models will get less accurate in rapidly change ecosystems (most of the world now?), because we are predicting to conditions outside the envelope of training data.
New study:
besjournals.onlinelibrary.wiley.com/doi/full/10....

12.08.2025 00:11 β€” πŸ‘ 2    πŸ” 1    πŸ’¬ 0    πŸ“Œ 0
Post image

Something to watch out for if you are using #chatGPT and #genAI for statistical advice: it doesn't ask enough questions. A good human statistician will ask lots of questions about context before it gives advice
ecoevorxiv.org/repository/v...

12.08.2025 00:07 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Quality R analysis with large language models 1-day workshop for using large language models to generate R code.

Want to know how to use #genAI to get good quality #rstats faster than typing? Check out my 1 day workshop, notes free here www.seascapemodels.org/R-llm-worksh... #chatGPT #largelanguagemodel

12.08.2025 00:04 β€” πŸ‘ 5    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

@bluecology is following 20 prominent accounts