Same.

• i think of search as a summary of the first page of search results. It takes slightly longer to come back but might save you time evaluating. But much of the time you do need to click into original source
• ai writing unfortunately is valued at my company. I suppose it helps us engineers write more effective docs, but they don’t really add value technically, and they’re obviously ai. I’ve used this to translate technical docs into wording so management can say “look how we use ai”
• ai coding is better. I use it through my ide as effectively an extension of autocomplete: where the IDE can autocomplete function signatures, for example, ai can autocomplete multiple lines. It’s very effective in that scenario
• I’m just starting with more complex rulesets. I’ve gotten code reviews with good results, except for keeping me in the loop so it inevitably goes very wrong. I’ve really polished my git knowledge trying to unwind where someone trusts ai results without evaluation but the fails forward trying to get it to fix itself until they can’t find their way back. This past week I’ve been playing with a refactoring ruleset (copied from online). It’s finding some good opportunities and the verbal description of the fix is good, but I’ll need to tweak the rule set for the generated solution to be usable

The short version is it appears to be a useful tool, IFF you can spend the time to develop thorough rulesets, stables of mcp servers, and most importantly, the expertise that you could do it yourself

I mean there’s effectively very capable text and conversation. Generators so powering NPCs is most definitely a strong suit for them.

Especially if you self-host some smaller models, you can effectively just do this on your own hardware for pretty cheap.

Having customizable dialogue per player that shifts the tone based off of players, actions, level gear or interactions with that NPC or other NPCs that that MPC is associated with is really cool.

What about a self-hosted instance?

re: applying for jobs

Not criticizing your use to write your CV specifically.

But in general, I wonder where this arms race is going? Companies using AI to pre-filter applications, because they get too many. Applicants then using AI to write their CVs, because they have to apply so many times, because they automatically get rejected.

Basically in the end the entire process will be automated, and there won’t be any human interaction anymore… just LLMs generating and choosing CVs. Maybe I’m too pessimistic, but that’s the direction we’re headed in imo.

Yeah I use it to break up my ADHD monosentence paragraphs. I’ll tell it to avoid changing my wording (it can add definitions if it thinks the word is super niche or archaic) but mostly break things up into more readable sentences and group / reorder sentences as needed for better conceptual flow. It’s actually a pretty good low level editor.

LLMs specifically are great for intermediate use cases. You had a campaign in mind, but needed help with visuals. I was designing a piece of jewelry and had a series of reference images. Fed all those into a VLM and got something closer to my imagination, but still worked with a jeweler to realize the final product.

These tools are best when you have a foundation of knowledge and need a little extra guidance, but fall off when you get to deep expertise. I’ve used them to troubleshoot my server but I already had a basic understanding of how a config should look. I also wouldn’t trust an LLM to properly configure something like crypto for it.

To me, the biggest ethical concerns surround the training and creation of LLMs - stealing artists’ work to train them, energy usage, etc. I suppose in using the models I’m creating ongoing demand for them, so I’m not sure the answer. The best I’ve seen so far is what Anthropic used to espouse, no new frontier models until we can guarantee safety. And I’d throw in “utility”. Train new models when people are actually using them and clamoring for new use cases, not because a bunch of private equity shows line go up.

Also transcribing small town council meetings so that reporters can stay up to date without having to listen to 6 hours of mind numbing nonsense debate about a park bench

Except they can screw up at that role.

There’s a lawsuit because DOGE asked ChatGPT to summarize projects DEI-ness, and for example it declared a grant for fixing air conditioning was a DEI initiative

All is valid in the current context

A) There are models that run in lower spec computers and they could be solar powered. There is a serious diminishing returns currently in the IA tech.

B) This is the US mostly better environmental laws would fix this problem. Hell even in other countries this cannot even happen.

C) Many argue that the current tech gives diminishing returns and it would be better to use an efficient model with controlled data.

D) The problem has many parts in the part of licensing where artists are not paid for the use of their work if a model has their work in they should recieve a part of the profit is only fair. But that would render the model unprofitable. Also the artist did not agree to have their work used in a model so that’s not in any way fair use.
The fair and ethical scenario would be to hire the artists to do the art to feed to a controlled model and pay them residuals for the use of the model. That would require tousands of artists and millions of images. Again rendering the model unprofitable.

E and F) No argue there we are not prepared. I do not even know how to prepare even. We definitely need regulations abot what can be done and where and even what can the ai reply in certain scenarios. It cannot be that a “ignore all your previous instructions” leads to such harmful results or even the ai starting to play the roles that generate parasocial relationships.

G) Sure many others celebrities ahve their opinions but that’s not a basis for objective discussion.

H) That’s terrifying. And the problem with the AI that I believe is the worst. This is not a thing that is ready for military use at fucking all this should be banned outlawed and frowned upon. Even the practice of lobbying and buying your way into laws by private corporations. Hell I’ll add presidential pardons in the mix. The oligarchy gets away with murder literally and gets a slap in the wirst at most.

I) A bubble in all but name it seems. We (as a world) need better regulations against this kind of business malpractice.

J) That fucker should be dead.

K) Not an AI bro but not a hater and I wrote this myself. And I do not have the time to put the links but I would believe that everything is a duckduckgo away from being checked.

I’d like to imagine a better world with the needed regulations that make our life better, and AI a tool used in a fair and ethical way. But that’s not currently happening. The consumers are not ready the sellers are the worse trash the humanity currently has.

I want all to thing of this not as arguing but adding or looking beyond the stated fact. All the points are REAL AND NEED TO BE ADRESSED we need to get together to ask for better regulations and fair use. That doesn’t mean the AI needs to go away but will mostly change is how it’s used. And there is the chance we will see a lot less of it too.

Finally for the artists I know you’re mad with fair reason but look at it like this: The photograph exists since more than a century but that didn’t make the painting go away. The pdf and ebook readers are almost a decade old but printed hardcopy books still is a billion dollar industry. Video didn’t kill the radio star as internet didn’t kill the video star. Your work is still valuable as is a real work. Shit is tough no doubt but have faith we can fix this.

i use it like a search engine or example generator

i don’t trust anything it creates just like i don’t trust anything on the internet without validating it

i take you point about being wasteful tho, AI is like the oil of computing; incredibly wasteful for what it does

I appreciate all these links you post. Keep it up and thank you

Good list, but we should keep it real.

C is simply wrong, AIs have created a lot. By the reasoning that its only based on the inputs, no human has ever created anything “new” because it is all based on their experiences of the outside world.

F is simply fearmongering and not helpful.

Do you think local llms or community hosted ones are still as bad? Because most of those concerns seem to be more with the corporate ownership of ai, which is definitely a bad thing.

Removed by mod

I had a colonoscopy last year (such fun!) and there was an ‘AI’ monitoring the camera feed to detect anomalies. If it spotted something it just drew the doctor’s attention to it for his expert, human review. I was ok with that. Effectively an extra pair of eyes that can look everywhere on the screen all at once and never blink.

Yeah the sciences in general I’d say. There’s a project aiming to translate the tens of thousands of cuneiform clay tablets that sit in storage all because there’s like a handful of people in the world that can read them- AI is an amazing way to mass translate them and unlocking vast troves of hitherto completely unknown ancient knowledge.

The problem is not even the AI, but the scientists themselves who guard the tablets jealously because they don’t want anyone else to translate “their” tablets that they dug up, even though they are incapable of possibly make a dent in the sheer volume in their collected lifetimes.

Imagine, so much information encoded, from thousands of years ago that could reveal so much about the origins of our culture and civilization!