iiinova — where your digital self keeps evolving

Sean Hamilton Palicki @seanhp.bsky.social · 14/09/2026

NEW PAPER in Social Science Computer Review on environmental costs of swapping task-specific ML models with LLMs for sentiment analysis, named entity recognition and text classification. Takeaway: Big models do not always outperform smaller ones, yet they may use up to 150,000x more energy. #Below2

120

Reposted by Sean Hamilton Palicki

The COMPTEXT Association @comptext.bsky.social · 25/04/2026

That's a wrap on #COMPTEXT2026! Huge thanks to lead organiser @chrisguarnold.bsky.social and IDAI/@unibirmingham.bsky.social team, plus all tutors, chairs, discussants & sponsors. Conference dinner at a world-renowned cricket stadium—what a treat. Safe travels home; see you at COMPTEXT 2027!

03515

Reposted by Sean Hamilton Palicki

Mark Chadbourn @chadbourn.bsky.social · 26/04/2026

Trump is so unlucky. Every time he’s trying to win a close election or is bombing in the polls because of his failed policies someone lets off a gun in his vicinity.

37737180

Sean Hamilton Palicki @seanhp.bsky.social · 05/03/2026

Software development in 2026 increasingly feels like giving up control. Move beyond "role-play and pray" with vibe-coding best practices.

000

Sean Hamilton Palicki @seanhp.bsky.social · 01/03/2026

Here's a totally harmless hammer to break ChatGPT.

000

Sean Hamilton Palicki @seanhp.bsky.social · 01/03/2026

To be critical of the government is the most American thing ever. Dario Amodei, PhD, the CEO of Anthropic has a spine for refusing to let our Mad King use Claude for domestic mass surveillance or fully autonomous weapons. I just bought a Claude Pro subscription. Fight on scholar. 🎓⚔️

010

Reposted by Sean Hamilton Palicki

Jana Bernhard-Harrer @janabernhard.bsky.social · 21/11/2025

In the latest #aBitOfCCS episode, we dig into research showing that LLMs often deliver only marginal gains at major environmental cost. 🌱🌍 Sean-Kelly Palicki breaks down when simpler methods beat large models—and why a CO₂-adjusted F1 score matters.🌱 Listen here ⤵️ aboutccs.net/abitofccs/

aboutccs.net

#aBitOfCCS – What is it about CCS

052

Reposted by Sean Hamilton Palicki

Johannes B. Gruber @jbgruber.bsky.social · 07/10/2025

Academic life hack: check which papers AI hallucinated most often and write them 🚀🚀🚀

0174

Sean Hamilton Palicki @seanhp.bsky.social · 10/10/2025

Speaking of Europe, I've been deGoogling on the grounds of data privacy, personal freedom, and surveillance capitalism. Yeah, it's kind of annoying but it's also pretty great. Here's what started my journey and how you can get started www.youtube.com/watch?v=u_Lx...

youtube.com

I'm DONE with Google

YouTube video by PewDiePie

010

Sean Hamilton Palicki @seanhp.bsky.social · 18/08/2025

📢 New Preprint: "Don’t Look Up: Evaluating the Tradeoff between Performance and Sustainability of LLMs for Text Analysis." Providing guidance for social scientists, we challenge the "bigger is better" narrative, instead encouraging social and environmental responsibility. osf.io/preprints/so...

095

Reposted by Sean Hamilton Palicki

Hacker News Top 100 Bot (Unofficial) @hn100.bsky.social · 19/05/2025

Germany drops opposition to nuclear power in rapprochement with France Discussion

001

Reposted by Sean Hamilton Palicki

Lisa Fazio @lkfazio.bsky.social · 11/05/2025

In case you missed it - last week they had the F1 drivers race giant lego cars for one lap and mayhem ensued. Some of the best racing drivers on the planet giggling like schoolboys as they race (and wreck) the cars 🤣 youtu.be/pdxiNBwxYxU?...

youtu.be

F1 LEGO Race In Full | 2025 Miami Grand Prix

YouTube video by FORMULA 1

174

Sean Hamilton Palicki @seanhp.bsky.social · 10/05/2025

Gossip on the internet, hang out, and let some political party decide what you’re going to do. And the next thing you know, you go, ‘Where did my health care go? Why are we fighting a war over here?’ Because you didn't pay attention. Maybe you should pay attention. www.wired.com/story/cyberp...

wired.com

A Founding Father of Cyberpunk Isn't Surprised By Its Comeback

Mike Pondsmith, who wrote the tabletop RPG that inspired 'Cyberpunk 2077,' explains why the genre feels vital in 2020.

010

Reposted by Sean Hamilton Palicki

Simon Willison @simonwillison.net · 06/05/2025

Looks like the original artisanal pencil sharpener was David Rees, here's his short film from twelve years ago vimeo.com/60718161

vimeo.com

HOW TO SHARPEN PENCILS

Official Selection: 2013 Austin Film Festival 2013 SF Docfest 2013 Sidewalk Film Festival (Winner: Best Documentary Short) 2013 Indie Memphis Film Festival (Winner:…

5281

Sean Hamilton Palicki @seanhp.bsky.social · 04/05/2025

May the 4th be with you! Watch Andor Season 2 ;)

000

Sean Hamilton Palicki @seanhp.bsky.social · 03/05/2025

This article is a great reminder to “appreciate” that we’re living in a cyberpunk future. www.noahpinion.blog/p/welcome-to...

noahpinion.blog

Welcome to The Future

This time it arrived right on schedule.

010

Sean Hamilton Palicki @seanhp.bsky.social · 01/05/2025

Please add me too! Thank you! 🙏🏼

100

Sean Hamilton Palicki @seanhp.bsky.social · 01/05/2025

I’m thinking Small Language Model + Whisper Speech-to-Text + RAG AI System that runs on an Inference Optimized Raspberry Pi 5. Let’s keep this implementation local and open 😉

000

Sean Hamilton Palicki @seanhp.bsky.social · 01/05/2025

Great read about the emergence of a topic that’s becoming increasingly part of popular culture: An AI enabled “afterlife”.

120

Reposted by Sean Hamilton Palicki

Clint Claessen @clint0475.bsky.social · 25/04/2025

All praise to @seanhp.bsky.social @isaacbravo.bsky.social for producing and presenting this fantastic paper! Check out the results and the R adaptation of CodeCarbon on linktr.ee/comptext25

linktr.ee

@comptext25 | Linktree

S. Palicki, I. Bravo & C. Claessen (2025)

1136

Sean Hamilton Palicki @seanhp.bsky.social · 26/04/2025

Thanks for giving us the chance to share!

020

Sean Hamilton Palicki @seanhp.bsky.social · 24/03/2025

See you at COMPTEXT 2025 in Vienna! I'll be presenting "Don’t Look Up: Evaluating the Tradeoff between Accuracy and Sustainability of LLMs for Text Analysis"

040

Reposted by Sean Hamilton Palicki

Social Media Lab @socialmedialab.ca · 20/02/2025

Things are changing fast. ICA, the largest academic conference for communication researchers, has begun omitting paper titles from acceptance letters to smooth out issues during the travel visa application process. #AcademicFreedon #BandAidSolution

0115

Sean Hamilton Palicki @seanhp.bsky.social · 20/02/2025

"Trump and his American oligarchs are now openly aligning themselves with Putin and his Russian oligarchs," Bernie Sanders wrote Monday on X. "This Putin-Trump alliance means abandoning our allies, supporting authoritarianism and undermining our democratic traditions."

163

Sean Hamilton Palicki @seanhp.bsky.social · 20/02/2025

Remember the Mueller Report? en.m.wikipedia.org/wiki/Mueller...

en.m.wikipedia.org

Mueller report - Wikipedia

000

Sean Hamilton Palicki @seanhp.bsky.social · 20/02/2025

Trump’s former vice-president, Mike Pence, who said in a social media post: “Mr. President, Ukraine did not ‘start’ this war. Russia launched an unprovoked and brutal invasion claiming hundreds of thousands of lives. The Road to Peace must be built on the Truth.” www.theguardian.com/world/2025/f...

theguardian.com

Trump calls Zelenskyy a dictator amid fears of irreconcilable rift

Remark follows Ukrainian leader’s claim US president living in a Russian ‘disinformation bubble’

230

Reposted by Sean Hamilton Palicki

Brian Merchant @bcmerchant.bsky.social · 18/02/2025

I've been writing BLOOD IN THE MACHINE for months, but I was on the fence about going all-in. Recent events have made the call easier. Elon Musk, the Silicon Valley oligarchs and their AI are taking over. I want to do all I can to push back. I hope you'll join me, and I'm asking for your support.

bloodinthemachine.com

The tech oligarchs and their AI are taking over. Let's fight back.

Alright, I’m really doing this thing. Support Blood in the Machine and fearless independent tech journalism.

241145365

Reposted by Sean Hamilton Palicki

Christoph Harig @charig.bsky.social · 19/02/2025

Brazil’s former President Bolsonaro charged over alleged coup that included a plan to poison Lula However, the prosecutor-general perpetuates the narrative that the armed forces had been a 'victim' of the plot, despite the obvious institutional involvement apnews.com/article/bols... #CivMilSky

apnews.com

Brazil’s former President Bolsonaro charged over alleged coup that included a plan to poison Lula

Brazil’s top prosecutor Paulo Gonet has formally charged former President Jair Bolsonaro with attempting a coup to stay in office after his 2022 election defeat.

28218

Reposted by Sean Hamilton Palicki

Joe @twitskeptic.sirvin.com · 18/02/2025

Let's see how the intrepid Elon inspired team at DOGE - the folks who are currently examining your tax records - are doing! Today they put up their their idea of "open data" which they call a "Wall of Receipts". And look here - two cancelled DEIA contracts for $9,999,999. Wow, they saved $20M! 1/

Two highlighted contracts for $9,999,999 each.

20683271

Sean Hamilton Palicki @seanhp.bsky.social · 18/02/2025

LLMs don’t do photography, nor do they investigate, think, or create (controversial). The Guardian being paid by OpenAI doesn’t mean OpenAI will be replacing the human there. Journalists, photographers, graphic designers, etc. need funding. OpenAI needs content. It’s symbiotic.

000

Sean Hamilton Palicki @seanhp.bsky.social · 18/02/2025

Language models don’t produce the news; they consume it. Why can’t a language model be considered a content distribution channel for a mainstream newspaper? The current paradigm of tech companies pirating journalism is clearly broken, but why are you against a licensing model?

100

Sean Hamilton Palicki @seanhp.bsky.social · 17/02/2025

From a tech perspective, the Groq platform is awesome for trying open models like Deepseek-R1, Mistral and Whisper. It's cheap, easy to access and it's stupid fast. Available via API and chat "playground" from US servers. groq.com/pricing/

groq.com

Groq is Fast AI Inference

Groq offers high-performance AI models & API access for developers. Get faster inference at lower cost than competitors. Explore use cases today!

000

Sean Hamilton Palicki @seanhp.bsky.social · 16/02/2025

The bge-m3 word embeddings work well for vector-based multilingual information retrieval. The easiest ways to implement this would be to use a vector database like Weaviate or Milvus. You can also use this to power a multilingual LLM RAG application using Ollama. huggingface.co/BAAI/bge-m3

huggingface.co

BAAI/bge-m3 · Hugging Face

We’re on a journey to advance and democratize artificial intelligence through open source and open science.

010

Sean Hamilton Palicki @seanhp.bsky.social · 14/10/2023

“However, while television is still bound to regulations that require a modicum of truthful content, social media is a battleground of facts, lies, and deception, where governments, journalists, law enforcement, and activists are on an uneven playing field.”

000

Reposted by Sean Hamilton Palicki

Zeke Hausfather @zekehausfather.com · 13/10/2023

There is increasing evidence that the world has warmed faster over the past 15 years than it has since the 1970s. Surface records, ocean heat content, and the Earth's energy imbalance all support an acceleration of warming, as I argue in today's @nytimes.com: www.nytimes.com/2023/10/13/o...

11551365

Reposted by Sean Hamilton Palicki

Stefanie Walter @stefaniewalter.bsky.social · 02/10/2023

What is the best sampling approach for multilingual texts? In our new paper in Computational Communication Research, we (Sean-Kelly Palicki, Wouter van Atteveldt, Alice Beazer, & Isaac Bravo) find that experts know best www.aup-online.com/content/jour... commsky polcomm polisky PolDataSky

12811

Sean Hamilton Palicki @seanhp.bsky.social · 28/08/2023

www.theatlantic.com/books/archiv... The clockwork bird can sing, but only the song with which it has been programmed. It can’t improvise. It can’t riff. It can’t surprise.

000

Sean Hamilton Palicki @seanhp.bsky.social · 09/08/2023

I’m interested in what you’re saying. I wonder if this is an inference based on previous comparable rule changes, or if this is just something that sounds reasonable. If it’s the former, then indie hackers and researchers beware. If it’s the latter, then I’d love to be directed to evidence.

100

Sean Hamilton Palicki @seanhp.bsky.social · 09/08/2023

Also, are we really going to pretend like corporations such as Google have the same legal treatment as everyone else. No way. As researchers or entrepreneurs, we should be aware that there are sharks in these waters and that we can’t afford a boat.

120

Sean Hamilton Palicki @seanhp.bsky.social · 09/08/2023

I’m no legal scholar, but I’m aware of several data scraping/fair use cases against Google, OpenAI and Github that are ongoing and related to training LLMs. The law hasn’t caught up to this problem yet, but we’ll probably know soon, and many other non-AI projects will probably be similarly impacted.

000

Sean Hamilton Palicki @seanhp.bsky.social · 09/08/2023

He gathered the data questionably, analyzed it privately, then sold the results. In the web scraping community we distinguish between gathering data for research and aggregate-level analysis and using the data for commercial gain. His site is a not-so-innocent bystander in the revolt against AI.

1170