Pour kettle and let steep the gods of tea. I built NewsBlur and Turn Touch.
1329 stories
·
871 followers

You’re Running on Autopilot Way More Often Than You Think

1 Comment

Think about what you did this morning. You woke up, brushed your teeth, made coffee, maybe scrolled your phone, maybe drove the same route to work that you do every day. How many of those things did you actually think about? According to new research, probably none. Scientists say nearly nine out of every ten daily actions happen on autopilot, with our brains running the show long before conscious thought gets involved.

The study, published in Psychology & Health, tracked 105 people for a week. Participants were pinged six times a day and had to report what they were doing, along with how deliberate or automatic it felt.

Across more than 3,700 reports, researchers found that 88 percent of behaviors were carried out automatically, while about two-thirds were triggered by habit rather than decision-making.

Lead researcher Amanda Rebar, an associate professor at the University of South Carolina, explained that this automation shows up in two ways.

“Habitual instigation occurs when environmental cues automatically trigger the decision to do something, like reaching for your phone when you hear a notification. Habitual execution happens when you perform an action smoothly without thinking about the mechanics, such as brushing your teeth or driving a familiar route,” she said in a statement.

Most people like to imagine themselves as rational actors, carefully weighing each choice they make. In practice, the study shows, life is closer to a string of well-worn loops. And those loops don’t vary much. Age, gender, and relationship status had no real effect on how habitual someone’s behavior looked.

One exception was exercise. People were more likely to start workouts based on cues, which could mean a reminder on their phone or a regular time of day, but still had to engage consciously once they got moving. Running, lifting, or cycling doesn’t complete itself, even if the decision to start feels automatic.

Habits, it turns out, often line up with what people want. Almost half of all reported behaviors were both intentional and automatic, while only a small fraction clashed with someone’s goals. That makes habits a surprisingly strong ally for anyone hoping to change.

Benjamin Gardner, a psychology professor at the University of Surrey and co-author of the study, said strategies for habit formation are more effective than willpower alone.

“For people who want to break their bad habits, simply telling them to ‘try harder’ isn’t enough,” he said. Building cues for healthier choices—or dismantling the ones tied to unhelpful patterns—might be the clearest path to change.

Most of what you do today will unfold without much thought. The trick, researchers suggest, is shaping those automatic moments to get you one step closer in the direction you actually want to go.

The post You’re Running on Autopilot Way More Often Than You Think appeared first on VICE.

Read the whole story
samuel
9 days ago
reply
I believe this applies to speech as well
Cambridge, Massachusetts
Share this story
Delete

ChatGPT Is Blowing Up Marriages as Spouses Use AI to Attack Their Partners

2 Comments

ChatGPT Is Blowing Up Marriages as Spouses Use AI to Attack Their Partners

Maggie Harrison Dupré for Futurism. It turns out having an always-available "marriage therapist" with a sycophantic instinct to always take your side is catastrophic for relationships.

The tension in the vehicle is palpable. The marriage has been on the rocks for months, and the wife in the passenger seat, who recently requested an official separation, has been asking her spouse not to fight with her in front of their kids. But as the family speeds down the roadway, the spouse in the driver’s seat pulls out a smartphone and starts quizzing ChatGPT’s Voice Mode about their relationship problems, feeding the chatbot leading prompts that result in the AI browbeating her wife in front of their preschool-aged children.

Tags: ai, generative-ai, chatgpt, llms, ai-ethics, ai-personality

Read the whole story
samuel
10 days ago
reply
I’ll quote the passage that I think this entire article boils down to, which is the final section:

The couple whose marriage had fallen apart after a previous reconciliation — the one in which the wife used ChatGPT to text their 10-year-old son about divorce — eventually agreed that splitting up was the right choice. As they prepared to file their case, things felt civil, the husband said.

Until one day, that is, when his wife broke a joint agreement to avoid large purchases before their case was filed. Feeling frustrated and betrayed, he pulled out his phone and started messaging with something he knew would respond: ChatGPT.

“I was just in rage,” said the husband. “And I kept engaging with ChatGPT that evening, and it kept telling me that this is a legal problem, and that she crossed a major line, and here’s how to bring it up with my lawyer, and here’s what the lawyer should file.”

The man described his vexation building as he continued to talk to the bot. ChatGPT, fed only his side of the story, characterized his wife’s behavior as manipulative, calculating, and reckless; her actions were deeply serious, it said, and encouraged the husband to take legal action.

The next day, distressed and still simmering with anger, the husband took the situation to his human lawyer. And as it turned out? It wasn’t a big deal at all.

“When I talked to my actual lawyer the next day, my lawyer was like, ‘that’s fine,'” the man recalled. “And at that point I realized — oh my god, I just went down the same spiral.”

“I can see how it happens,” he said. “It happened firsthand to me.”
Cambridge, Massachusetts
Share this story
Delete
1 public comment
rocketo
9 days ago
reply
really fucking sad
seattle, wa

I think "agent" may finally have a widely enough agreed upon definition to be useful jargon now

1 Comment

I've noticed something interesting over the past few weeks: I've started using the term "agent" in conversations where I don't feel the need to then define it, roll my eyes or wrap it in scare quotes.

This is a big piece of personal character development for me!

Moving forward, when I talk about agents I'm going to use this:

An LLM agent runs tools in a loop to achieve a goal.

I've been very hesitant to use the term "agent" for meaningful communication over the last couple of years. It felt to me like the ultimate in buzzword bingo - everyone was talking about agents, but if you quizzed them everyone seemed to hold a different mental model of what they actually were.

I even started collecting definitions in my agent-definitions tag, including crowdsourcing 211 definitions on Twitter and attempting to summarize and group them with Gemini (I got 13 groups).

Jargon terms are only useful if you can be confident that the people you are talking to share the same definition! If they don't then communication becomes less effective - you can waste time passionately discussing entirely different concepts.

It turns out this is not a new problem. In 1994's Intelligent Agents: Theory and Practice Michael Wooldridge wrote:

Carl Hewitt recently remarked that the question what is an agent? is embarrassing for the agent-based computing community in just the same way that the question what is intelligence? is embarrassing for the mainstream AI community. The problem is that although the term is widely used, by many people working in closely related areas, it defies attempts to produce a single universally accepted definition.

So long as agents lack a commonly shared definition, using the term reduces rather than increases the clarity of a conversation.

In the AI engineering space I think we may finally have settled on a widely enough accepted definition that we can now have productive conversations about them.

Tools in a loop to achieve a goal

An LLM agent runs tools in a loop to achieve a goal. Let's break that down.

The "tools in a loop" definition has been popular for a while - Anthropic in particular have settled on that one. This is the pattern baked into many LLM APIs as tools or function calls - the LLM is given the ability to request actions to be executed by its harness, and the outcome of those tools is fed back into the model so it can continue to reason through and solve the given problem.

"To achieve a goal" reflects that these are not infinite loops - there is a stopping condition.

I debated whether to specify "... a goal set by a user". I decided that's not a necessary part of this definition: we already have sub-agent patterns where another LLM sets the goal (see Claude Code and Claude Research).

There remains an almost unlimited set of alternative definitions: if you talk to people outside of the technical field of building with LLMs you're still likely to encounter travel agent analogies or employee replacements or excitable use of the word "autonomous". In those contexts it's important to clarify the definition they are using in order to have a productive conversation.

But from now on, if a technical implementer tells me they are building an "agent" I'm going to assume they mean they are wiring up tools to an LLM in order to achieve goals using those tools in a bounded loop.

Some people might insist that agents have a memory. The "tools in a loop" model has a fundamental form of memory baked in: those tool calls are constructed as part of a conversation with the model, and the previous steps in that conversation provide short-term memory that's essential for achieving the current specified goal.

If you want long-term memory the most promising way to implement it is with an extra set of tools!

Agents as human replacements is my least favorite definition

If you talk to non-technical business folk you may encounter a depressingly common alternative definition: agents as replacements for human staff. This often takes the form of "customer support agents", but you'll also see cases where people assume that there should be marketing agents, sales agents, accounting agents and more.

If someone surveys Fortune 500s about their "agent strategy" there's a good chance that's what is being implied. Good luck getting a clear, distinct answer from them to the question "what is an agent?" though!

This category of agent remains science fiction. If your agent strategy is to replace your human staff with some fuzzily defined AI system (most likely a system prompt and a collection of tools under the hood) you're going to end up sorely disappointed.

That's because there's one key feature that remains unique to human staff: accountability. A human can take responsibility for its action and learn from its mistakes. Putting an AI agent on a performance improvement plan makes no sense at all!

Amusingly enough, humans also have agency. They can form their own goals and intentions and act autonomously to achieve them - while taking accountability for those decisions. Despite the name, AI agents can do nothing of the sort.

This legendary 1979 IBM training slide says everything we need to know:

A computer can never be held accountable. Therefore a computer must never make a management decision

OpenAI need to get their story straight

The single biggest source of agent definition confusion I'm aware of is OpenAI themselves.

OpenAI CEO Sam Altman is fond of calling agents "AI systems that can do work for you independently".

Back in July OpenAI launched a product feature called "ChatGPT agent" which is actually a browser automation system - toggle that option on in ChatGPT and it can launch a real web browser and use it to interact with web pages directly.

And in March OpenAI launched an Agents SDK with libraries in Python (openai-agents) and JavaScript (@openai/agents). This one is a much closer fit to the "tools in a loop" idea.

It may be too late for OpenAI to unify their definitions at this point. I'm going to ignore their various other definitions and stick with tools in a loop!

Tags: ai, generative-ai, llms, ai-agents, agent-definitions

Read the whole story
samuel
13 days ago
reply
Tools in a loop it is
Cambridge, Massachusetts
Share this story
Delete

Beyond Vibe Coding

2 Comments

Beyond Vibe Coding

Back in May I wrote Two publishers and three authors fail to understand what “vibe coding” means where I called out the authors of two forthcoming books on "vibe coding" for abusing that term to refer to all forms of AI-assisted development, when Not all AI-assisted programming is vibe coding based on the original Karpathy definition.

I'll be honest: I don't feel great about that post. I made an example of those two books to push my own agenda of encouraging "vibe coding" to avoid semantic diffusion but it felt (and feels) a bit mean.

... but maybe it had an effect? I recently spotted that Addy Osmani's book "Vibe Coding: The Future of Programming" has a new title, it's now called "Beyond Vibe Coding: From Coder to AI-Era Developer".

This title is so much better. Setting aside my earlier opinions, this positioning as a book to help people go beyond vibe coding and use LLMs as part of a professional engineering practice is a really great hook!

From Addy's new description of the book:

Vibe coding was never meant to describe all AI-assisted coding. It's a specific approach where you don't read the AI's code before running it. There's much more to consider beyond the prototype for production systems. [...]

AI-assisted engineering is a more structured approach that combines the creativity of vibe coding with the rigor of traditional engineering practices. It involves specs, rigor and emphasizes collaboration between human developers and AI tools, ensuring that the final product is not only functional but also maintainable and secure.

Amazon lists it as releasing on September 23rd. I'm looking forward to it.

O'Reilly book cover: Beyond Vibe Coding: From Coder to AI-Era Developer, by Addy Osmani. Features two hummingbirds, presumably because their wings vibrate!

Tags: books, oreilly, ai, generative-ai, llms, ai-assisted-programming, addy-osmani, vibe-coding

Read the whole story
samuel
27 days ago
reply
Whoa, this is the same Addy Osmani who wrote the O'Reilly Backbone.js book back when Backbone.js was everywhere (and continues to run NewsBlur).
Cambridge, Massachusetts
Share this story
Delete
1 public comment
rosskarchner
10 days ago
reply
My copy of this book arrived today-- I'm excited to read it, but it also seems like a 200ish page O'Reilly book with a $69 cover price is pretty emblematic of the bubble...

(I got it for $47 on thriftbooks.com)

Jack Osbourne Fires Back at Roger Waters For Insensitive Comments About Ozzy

1 Comment

Former Pink Floyd co-founder Roger Waters is under fire from Jack Osbourne after Waters shared a less than savory opinion about his late father, Ozzy Osbourne. The Black Sabbath frontman passed away on July 22 following a celebratory farewell concert, and Waters has since implied that he disapproved of Osbourne’s long legacy in rock and roll.

Waters’ comments came during an appearance on The Independent Ink podcast. His argument was that pop culture and celebrity often distract citizens from serious political issues. Taking on the opposing point of view for argument’s sake, he proposed how those in power might use pop culture to their advantage.

“‘How can we push this to one side? I know how to do it! We’ll do it with Taylor Swift or bubble gum or Kim Kardashian’s bum,'” he said. “Or Ozzy Osbourne, who just died, bless him, in his, whatever that state that he was in his whole life, we’ll never know. Although, he was all over the TV for hundreds of years with his idiocy and nonsense.”

Jack Osbourne Takes offense to Roger Waters’ comments about his late father ozzy

That wasn’t all he said, however. Waters then proceeded to make his comments about Ozzy Osbourne a bit personal.

“The music, I have no idea, I couldn’t give a fuck. I don’t care about Black Sabbath, I never did, I have no interest in … ‘Wahhhh!!!'” said Waters, sticking his tongue out and apparently imitating a Black Sabbath sound, before continuing, “and biting the heads off chickens or whatever they do. I couldn’t care less.”

Waters was even more disturbed when he was informed that Ozzy Osbourne actually bit the head off a bat (accidentally, thinking it was fake) and not a chicken. “Oh my God, that’s even worse, isn’t it?” he exclaimed. “I don’t know, is it worse to bite the head off a bat or a chicken?”

In response, Jack Osbourne stuck up for his late father by posting a comment in an Instagram story. “Hey Roger Waters – fuck you. How pathetic and out of touch you’ve become. The only way you seem to get attention these days is by vomiting out bullshit in the press. My father always thought you were a cunt – thanks for proving him right,” he wrote.

Photo by Larry Busacca/Getty Images for Tribeca Film Festival

The post Jack Osbourne Fires Back at Roger Waters For Insensitive Comments About Ozzy appeared first on VICE.

Read the whole story
samuel
29 days ago
reply
What good is there in speaking ill of the dead. Shared for the family's final word on the matter.
Cambridge, Massachusetts
Share this story
Delete

sosumi.ai: Apple Developer Docs for LLMs

1 Comment

Sosumi.ai:

Ever notice Claude struggling to write Swift code? It might not be their fault!

Apple Developer docs are locked behind JavaScript, making them invisible to most LLMs. If they try to fetch it, all they see is “This page requires JavaScript. Please turn on JavaScript in your browser and refresh the page to view its content.

This service translates Apple Developer documentation pages into AI-friendly Markdown.

Perfect little audio easter egg on the page. Beautiful Markdown output too. Look at my boy, all grown up, teaching robots how to program.

I do regret, though, that I didn’t define or influence the fenced style for code blocks. If I had, instead of this:

```swift
// An array of 'Int' elements
let oddNumbers = [1, 3, 5, 7, 9, 11, 13, 15]
```

You could do this, which looks so much better:

``` Swift:
// An array of 'Int' elements
let oddNumbers = [1, 3, 5, 7, 9, 11, 13, 15]
```

Those all-lowercase language identifiers, with no preceding space, just look a little lazy. I realize why GitHub’s ```-fenced code blocks took off (they’re the only code block style most Markdown users know, I suspect), but they don’t look nearly as nice, to human readers, as my original tab-indented style.

Read the whole story
samuel
30 days ago
reply
The tab-indented code block style means reformatting all of your code. The triple backtick means you can simply wrap code.

I get why he made that mistake originally, but now that it's been decades, why doesn't he recognize the superior delineator?
Cambridge, Massachusetts
zwol
30 days ago
I have never heard John Gruber *ever* admit that he made a mistake
Ailuridae
17 days ago
and yet he made so many with MD (I mean it's one of the best for a lot of things, I purposely picked it for my extensive personal notes, but damn it has a lot of problems, only some of which are fixed with CommonMark)
Share this story
Delete
Next Page of Stories