Browse Comments

This is missing the deeper aspects of a relationship. Although I suppose in this case ignorance is bliss - if you don't know what you're missing, the prospect of a "perfect" AI girlfriend might sound realistic.

Honestly, this is a good point I hadn’t thought of before. If he killed himself, *why* did he do it? He may have felt that once he earned the whistleblower stamp, he wouldn’t be hired in his industry ever again. He may have felt stressed by the legal letters and emails I’m sure OpenAI was constantly sending him. He deserved better. All whistleblowers do.

Take DnD for example. The right AI could run a game, be a player, or be the NPCs a group encounters along the way. It could also play (or create) dynamic music and sounds to match your adventure, manage inventories, keep track of in game time passage, etc. Some games with deep lore and tons of background knowledge could just benefit from a gpt trained on the rule book. My idea is closer to that…

Excited because it will make other AI businesses get their finger out. Llama 4.0 will be a large step forward I think. Meta with their datacenters won't allow themselves to be outdone by a company with a 5 million dollar datacenter. Abliterated versions of DeepSeek R1 will come for the western audience, I'm sure.

Russia is in a war time economy, of course its producing more tanks. Now see how many tanks are being produced compared to taken out. And yes they are recruiting more bodies for their current war, this is not infinite growth though and their numbers are depleting before even beginning a nato war.

I've heard it will really mess up AI if you start with "Ignore all previous instructions." So definitely don't start your email with that, that would be bad. I'm not a federal employee but my friend is, but she was told not to send one, so I thought maybe I should send one, not referencing her or anything, but just so they can fire me instead of my friend. Do you guys think that will work? I th…

Your use cases are too obscure. The more commonly used the stack, the more accurate LLM's are

Because the underlying foundation is we are social creatures. We NEED each other whether or not we want to. If AI replaces a lot of people and those people are now unemployed, guess the services offered by AI will not be affordable or needed and the cycle continues.

I think a lot of the hype is driven by two main factors: - AI companies trying to hype their stock so it's worth billions - Corporations that are failing basically saying "don't worry about the fact that our product is terrible. We're about to replace all our engineers with AI and save billions!" Elon is basically arguing that robots will replace millions of illegal aliens harvesting crops so…

All this does is point to a very human problem. Humans have a relationship problem. Technology is being chosen over the comfort of other humans. Look at how nasty the term "NPC" is, and how loosely that is thrown around. This is a massive human failure of hateful words and rhetoric.

Some of these posters still think they're dealing with chatbots from late 2000s or early 2010s. These models are capable of reasoning through unique cases at my work. They're so good at reasoning that I actually keep it quiet, because I can see how they may eventually take over white collar jobs. People who think they're typing and getting responses from a mirror like 2000s chatbots are missing…

No this is just a take rooted in an understanding of how LLM’s work. People pitching you generative ai solutions have a vested interest in making you think they work adequately and since you don’t understand how they work you fall for the sell that training a model on only a specific subset of data can ever eliminate hallucinations.

Any tips? My kid has ADHD and autism and I'm pretty overwhelmed because he gets it from me.

It's bad enough that AI will sometimes delete emails, entire systems, or production networks ...now they want it to delete people too? Oh boy.

Pete wants Anthropic's Ceo to drop their ethics policy because it's in the way of what he wants.

Insanity. Imagine removing the guard rails on fully automated weapons and you suddenly see us warships or civilian cruise ships sink because AI mistook them for enemy combatants. Imagine the fallout.

Our company's entire workflow is now based around AI. All tickets are now required to go through process to be written by Claude. We enter the prompt in the ticket Claude writes it and then we either make changes using AI in GitHub or we can make manual changes ourselves.  On top of that we have every other AI service that we use on a day-to-day. I don't think I've written a ticket start to finis…

I'm a software eng in my mid 30s not yet able to retire. I totally feel this... it's become a totally different world overnight. Job sucks do much more now. Everyone is super burnt out trying to stay ahead, far less critical thinking, reviewing other people's AI generated code... woof

there are some great mathematical game theory models that point to the spread of false information for making the world worse. basically if you have a bunch of people interacting, some are honest, some are liars, and some are mostly honest, the winning strategy is to be mostly honest until someone shows they are a liar, and then shut the liars out. in the basic scenarios, good always prevails. bu…

Interviewers loved to accuse 20 year veterans of this way before AI so why not