iiinova — where your digital self keeps evolving

Andreea @aerins.bsky.social · 29/03/2025

Are web extensions on the rise again?

Reposted by Andreea

Hyperplane @hyperplane.bsky.social · 24/03/2025

We're launching an ebook on 28th this week 😳 Regular speech-to-text tech struggles with kids’ voices since they’re higher-pitched and less predictable. We worked on that by creating a smaller, more accurate model that works well with children’s speech, even in noisy or low-power settings.

111

Reposted by Andreea

Hyperplane @hyperplane.bsky.social · 24/03/2025

The principles of an indexing pipeline: 1/7

111

Andreea @aerins.bsky.social · 19/03/2025

Did anybody create a voice-only chatbot? What improvements would be cool from a UX perspective?

100

Reposted by Andreea

Hyperplane @hyperplane.bsky.social · 26/02/2025

"prompt engineering" is just fancy copy-pasting at this point people tweaking prompts like they're adjusting a car mirror, thinking it'll make them drive better you’re optimizing nothing, you’re just guessing

132

Andreea @aerins.bsky.social · 24/02/2025

Did anybody figure out how to prompt chat gpt to get the perfect animation using @gsap-greensock.bsky.social ? I am a beginner with animations and debugging them is not my forte :(

000

Reposted by Andreea

Prof Julia Steinberger 🌹🌱🌍 #ClimateAction #Fightfascism @jksteinberger.bsky.social · 23/02/2025

As promised, here are the slides I shared with students to convince them to NOT use chatGPT and other artificial stupidity. TL;DR? AI is evil, unsustainable and stupid, and I'd much rather they use their own brains, make their own mistakes, and actually learn something. 🪄

NO CHATGPT Or other artificial stupidity: motivation
First, clarity on distinguishing AIs:
Non-generative: grammar aid, translation, dictionary, text-to-audio (e.g. Natural Reader): no problem
As long as you use the appropriate tools (least intensive in data and server energy use).
Why? Because you provide the content. Your brain is doing the most important work
Generative: ChatGPT & Co.
You only supply the prompt, the AI supplies the content.
Why is this delegation of work problematic?
3 domains: ethical, environmental, intellectual engagement.

(Caveat: generative is probably ok for computer programming, where it can be useful and save time. Not relevant to this class.)

1) AI and ethics
Mass theft of all and everything
«learning» on books, articles, blogs, social media, images, music, cultural production, without permission of authors/creators, and leading to their mass joblessness. Profits are not reditributed to originators.
Permanent destruction of the mental health of underpaid precarious tech workers in the Global South (Kenya, Philippines …):
«correction» to avoid production of violent and pedophile contents etc, tech workers are obliged to watch and correct super violent contents for days on end, leading to extreme psychological suffering and trauma, from which recovery is doubtful. No or little compensation (certainly not at the level of the suffering inflicted).
In short, an industry built on theft of real human creation and sacrifice of real human health, profiting a few megafortunes.

2) AI and (un)sustainability

Massive consumption of electricity, water, server capacity for generative AI.
Outcome: keep fossil fuel companies in business, using up new renewable capacity, without any satisfaction of basic human needs.
Massive misappropriation of the finance necessary for climate and ecological action (renewable generation, efficiency and retrofit for buildings, public transit, infrastructures for cycling etc) towards AI industry.
Overall: undermine climate action, reinforce fossil industry, waste resources necessary for human development.

3) AI and intellectual engagement

First, what learning is (or should be) about:
The goal should not (only) be the reproduction of «correct» knowledge,
But mainly personal engagement and experience of thinking about topics of interest. Personal engagement = using one’s own brain.
The most important activity for learning and intellectual engagement is the experience of making one’s own mistakes, by trial and error, corrections based on new ideas, starting over again. Learning to recognise nuances, knowledge gaps, better explanations
This kind of learning is possible only through using your own brain, not AI.
Also, Ais are not «intelligent». At all.
They simply reproduce pre-existing patterns. They «bullshit», invent false references, false facts, false data, simply because those sound plausible. VERY DANGEROUS.
If you learn how to NOT use AI, and how to research facts and data on your own, this will serve you and your communities for the rest of your life.

23556172098

Reposted by Andreea

Katerina Marchán 🐈 @zkat.tech · 21/02/2025

(Boosts encouraged!) Hi! 👋 I’m Kat. I’ve been a professional dev for over 15 years. I maintained the NPM CLI for 5 of those. I sat in TC39. I work at Microsoft. I’m self taught w/a film degree and no formal CS edu. Are you new to tech? What do you wish you could ask some1 like me? Literally AMA

28421199

Reposted by Andreea

Hyperplane @hyperplane.bsky.social · 18/02/2025

Indexing messy PDFs? Try these: LayoutAnalyzer for messy formatting. TextDensityAnalyzer to tell text from images. VisualElementAnalyzer for charts/diagrams. TableDetector to grab tables. Convert them to images and use Vision Models like ColPali/ColQwen2.

121

Reposted by Andreea

JSHeroes @jsheroes.io · 07/02/2025

📢 We’re thrilled to announce the selected speakers from this year’s call for papers! 🔎 Check out the full lineup here: jsheroes.io/#speakers And make sure you join us in Cluj this May 💙 #JSHeroes #TechConference #TechInnovation #TechInsights #WebDev #JavaScriptTrends

1142

Andreea @aerins.bsky.social · 01/02/2025

Hello everybody! I launched a free newsletter where we deep dive into some of the toughest challenges for AI in production—especially now as POCs are everywhere, but real-world scalability gets the short end of the stick. I’d love for you to check it out! mlvanguards.substack.com

mlvanguards.substack.com

ML Vanguards | Substack

Escaping PoC purgatory: Your Weekly Guide to production paradise. Click to read ML Vanguards, a Substack publication.

021

Reposted by Andreea

Nicholas C. Zakas @humanwhocodes.com · 17/01/2025

Finishing up ESLint’s financials for 2024 and…it’s not pretty. ESLint ran at a nearly $60,000 deficit. We managed due to our prioritizing building up our reserves for several years prior in case of a rainy day. We can use your help closing the gap. eslint.org/donate

eslint.org

Donate - ESLint - Pluggable JavaScript Linter

A pluggable and configurable linter tool for identifying and reporting on patterns in JavaScript. Maintain your code quality with ease.

45524