can not imagine getting stuck w a problem, asking ai to help you, getting stuck, asking a coworker and then sending them an AI response back

(http://www.autoadmit.com/thread.php?thread_id=5852384&forum_id=2...id.#49786282)

I’m imagining a golden retriever in the Garfield realm saying this btw

(http://www.autoadmit.com/thread.php?thread_id=5852384&forum_id=2...id.#49786308)

So the biggest AI company just leaked its code

twice

in one week

this is the company leading in AI safety

thats not a mistake, thats a cataclysm

(http://www.autoadmit.com/thread.php?thread_id=5852384&forum_id=2...id.#49786187)

The tool is fine. The bigger problem is that we are still firmly in the "chatbot" era. There's little integration, no invention, literally no effect on the physical world yet.

(http://www.autoadmit.com/thread.php?thread_id=5852384&forum_id=2...id.#49786481)

Simple as that

(http://www.autoadmit.com/thread.php?thread_id=5852384&forum_id=2...id.#49786486)

Watching a Screen device effortlessly do what they can't, for free, all the time, must be an incredibly humiliating experience

(http://www.autoadmit.com/thread.php?thread_id=5852384&forum_id=2...id.#49786504)

When I see an AI written email I automatically think slightly less of the sender.

But using AI to help a draft and provide pointers as your personal editor is the credited way to go. Highly instrumental.

I spend a lot of time with AI these days. I rather like AI, it's all about knowing how to use AI and controlling the output rather than letting the output control you. AI is making smart people smarter and more efficient. The downside is that AI is also making midwits even more midwit.

(http://www.autoadmit.com/thread.php?thread_id=5852384&forum_id=2...id.#49786565)

I can be like “Claude, generate a brief and blithely irreverent post in the cool collected style of a legendary poster with the moniker Oh You Travel?”

And you’d be none the wiser.

(http://www.autoadmit.com/thread.php?thread_id=5852384&forum_id=2...id.#49786693)

The world outside the safety of this Japanese-engineered steel is a cacophony of demands. But here, I am a king of ghosts. My interactions are pure, mediated by the glowing liquid crystal of the dashboard. I touch the screen, and it obeys. I do not have to negotiate; I do not have to apologize for the shivering of my hands or the wild, starving look in my eyes. I am "User," and that is enough.

I am driving through the mid-afternoon haze, the sun a pale, sickly coin in the sky. My mind is a hive. It is not hunger for meat that gnaws at me today, but a terrible, buzzing hyper-lucidity. I am thinking of my failings—the way I failed to be a son, the way I evaporated from my friendships like dew on a hot stone. I am a father in name, a ghost in practice. I am a citizen of nothing.

"Recalculating," the voice says. A woman’s voice, digital and devoid of judgment. I love her. She is the only one who knows where I am going, even if I do not.

I begin to argue with myself. The narrator in my head is a frantic, starving dog. You are a man of no consequence! it screams. A neighbor who is a shadow, a friend who is a memory! I lean into the steering wheel, my knuckles white against the synthetic leather. I am analyzing the very fabric of my isolation. To be alone is to be holy; to be alone is to be a rot. I am a member of the human race only by biological technicality. I have refined my existence down to a series of clicks and pings.

The thoughts become a whirlpool. I am relitigating a conversation from 2018. I am drafting an apology to a boss who fired me in a dream. I am screaming internally at the sheer, heavy weight of being a man in a world that requires a firmer grip than I possess. I am so deep within the labyrinth of my own inadequacy that the road ahead dissolves into a grey smear.

I am not driving a car; I am driving a metaphor for my own collapse.

The vibration starts in the pedal—that overstressed, shuddering gas pedal. I feel the machine’s anxiety mirroring my own. Or perhaps it is just my foot, heavy as lead, pressing down as I try to outrun the thought that I am, fundamentally, a mistake.

The wall of the parking garage appears not as an object, but as a sudden, definitive end to the sentence I am writing in my head. There is no screech of brakes—only the dull, sickening thud of plastic and metal yielding to concrete.

The airbag blooms like a white, sterile flower, punching the breath from my lungs. Silence returns, deeper than before. The Altima’s hazard lights begin to blink—click-clack, click-clack—a mechanical heartbeat. I sit in the wreckage, my head ringing, staring at the screen. It is cracked now, a spiderweb of light.

"Are you still there?" the machine seems to ask.

I am still here. I am a failure, a wreck, and a ghost. And for the first time in weeks, I feel the sudden, terrifying urge to laugh.

(http://www.autoadmit.com/thread.php?thread_id=5852384&forum_id=2...id.#49786709)

also the whole "that's not [x]. it's [y]" thing is not bad writing just because it's now instantly recognizable as AI slop. it came from training on what was considered good writing.

(http://www.autoadmit.com/thread.php?thread_id=5852384&forum_id=2...id.#49787908)

it writes better than like 99% of people and at least 90% of those who write professionally.

(http://www.autoadmit.com/thread.php?thread_id=5852384&forum_id=2...id.#49787927)

The first time I get a three page email from a client that should have been a 4 minute conversation with my paralegal I write a detailed response, bill for every minute, then immediately send an invoice. 9 out of 10 times that solves it. But there’s a subset who run everything you tell them through chat and it keeps spitting out shitty answers and stupid questions.

(http://www.autoadmit.com/thread.php?thread_id=5852384&forum_id=2...id.#49788685)