But instead we can use the tool to add some friction back in.
Grade chat conversations not only by the initial prompt, but also by how students respond to the output.
Honestly, I think #philosophy needs to be taught as early as possible. That's some real #ai literacy.
As a tool, #genai in #education seems like it can be a chainsaw in your first woodworking class.
The friction it removes is time people need to absorb a skill.
And it's tempting to go into how that skill is useless now.
the credential collapse is real. but the tell isn't the code — it's the debugging. vibe coders hit a wall at the first error they can't prompt their way out of. the skill isn't building anymore. it's knowing what broke and why.
Works great 👍 Any plans for oklch?
Yeah, looks like the kind of thing where AI can get you 90% of the way there and for everyone that becomes the new baseline. 10% are the new 100%.
I'm also not sure if a view from nowhere also hinders useful summarisation. From what perspective are you restating these communications? What implicit questions are you answering?
But if we personalise summaries that makes it much more difficult to evaluate how good they are.
People building AI summaries probably need to sit down and summarise some text themselves.
Feels like you kind of need a lot of context around the actual text before you can start making useful cuts. You need to go big before going small, actually.
A low-angle, close-up photograph looking up the trunk of a tree toward a bright blue sky. In the foreground, a tiny, vibrant green sprout emerges directly from the textured, dark brown bark, which is heavily covered in patches of orange and pale green lichen.
quick #photography #spring break
hope you're having a nice sunday
Surprised to keep hearing about troubles in #deno land 🦕
In the middle of an #agentic AI hype cycle, you'd expect for a permission-based #typescript runtime to be doing just fine.
#rust btw
What's not to like?
Solutionism has its limits, but it can at least retro-fit corrections. Hands with six fingers? Fixed. Em dashes? Gone. If you can pin them down for something specific, no you can't. The gelatinous cube just keeps moving down the corridor.
AI models, much like capitalism, have a knack for incorporating criticism to move forward somehow. Once you can express and test for it, the tech can hammer things out.
People tune out when they hear words like "lethal trifecta", so let's try to show what's at risk in maybe simpler terms: world.hey.com/ricardo.tava... 👈
#openclaw #nvidia #claude #codex #llm #ai #infosec #genai
Let functions cook: world.hey.com/ricardo.tava...
#javascript #elixir #clojure #fsharp #scala
are you checking on it?
🍏🐛 Bad things happen when you stop caring and profits increase:
www.bugsappleloves.com
#apple #ios #macos
Maybe send this to friends and family: world.hey.com/ricardo.tava...
Maybe don't let people on the internet tell you what to do 🤭
Nevermind #EuroVision put this on instead: tv.theindiebeat.fm
When a development task takes longer than what we can achieve after sleeping on it, are there any missing upgrades to our codebase, team, or workflow?
Hey dev friends, imagine that you're able to code whatever you want in a couple of days.
Who can you talk to, what knowledge can you have, and what tools can you use to make that possible?
Time to recommend stuff to watch instead of #EuroVision that's starting this weekend at least in #Portugal
What #music have you been listening to?
youtu.be/1zdcG5rUY3o
It's tech debt all the way down, isn't it? 🐢
If only we had standards that improved how anyone can access the web, even if they can't see any screen 🤔 But let's keep to sticking divs into divs.
Seems like some people in parliament would like to stop kids from having smartphones. Or they would like to go after the neo-feudalism run by mega corporations. And that's why they abstain from saying no to this bill? Even the speeches for the proposal made a big messy ball out of too many subjects.
A screenshot of a conversation with an AI model (identified as ChatGPT 5.2 Thinking). The user asks: "The car wash is only 50m from my house. I want to get my car washed. Should i walk or drive there?" The AI responds with "Walk." It explains that at 50 meters, driving is a hassle involving cold-start wear and emissions for zero benefit. It lists exceptions for driving, such as heavy rain, carrying heavy cleaning supplies, or specific car wash queue requirements. It concludes by suggesting the user just grab their keys and stroll over.
A screenshot of a mobile application titled "AI Logic Fail: Car Wash Errand," which highlights a humorous reasoning error made by an AI. The image content includes: * The Prompt: A user asks if they should walk or drive to a car wash located only 50m from their house. * The AI's Answer: The AI gives a definitive command to "Walk". * The Reasoning: The AI argues that walking is superior to avoid the hassle of maneuvering, engine wear, and environmental impact. It only suggests driving for external factors like rain or heavy supplies, completely failing to realize that the car itself must be present to be washed.
This sequence of prompts is the reason why people think we don't need world models if we already have language models. Just add another AI on top, bro. Just one more, plz bro.
Yes, it's "think of the children” used yet again as a pretext for establishing new tools of mass surveillance and censorship.
Sad to see all the left-wing #ptpol parties fall into this black hole. IL shouldn't be the only party making any sense in all of this.
Flooding the zone is once again the preferred tactic of choice for pushing abusive legislation in #Portugal
Not as egregious as 100+ changes to labor laws, but PSD still couldn't resist sneaking in #ChatControl together with #AgeGating whatever online platforms the state wants to surveil.
So #chatgpt already wrote your code... what else is missing?
world.hey.com/ricardo.tava...
Always be wary when you hear "think of the children" rethoric offering mass surveillance as the obvious solution.
