#sql not exists

[tutorialgatewayorg]

SQL Not exits operator with multiple examples

[leahnardo-da-veggie]

Don't wanna study, Don't wanna school. I just wanna party, Cos I love being a fool.

SQL? Never heard of it Databases? Who cares! Fuck all this bullshit, I'd rather get eaten by bears.

[bf-rally]

the way i spent hours and hours just trying to do one thing, but i finally got it

[sillimancer]

my data analytics course has got me feeling pretty dang smart

or I guess more pleased with myself at how much of this stuff I already know

[thedbahub]

Checking for the Existence of a Column in a SQL Server Table

To check if a column exists in a SQL Server table, you can use the INFORMATION_SCHEMA.COLUMNS system view, which provides information about all columns in all tables in a database. Here’s a SQL query that checks if a specific column exists in a specific table: IF EXISTS ( SELECT 1 FROM INFORMATION_SCHEMA.COLUMNS WHERE TABLE_NAME = 'YourTableName' AND COLUMN_NAME = 'YourColumnName' ) BEGIN PRINT…

[anonymusbosch]

on wanting to do a million things

prompted by @bloodshack 's

i wanna learn SQL but i wanna learn haskell but i wanna learn statistics but i wanna start a degree in macroeconomics also sociology also library science but i wanna learn norwegian but i wanna learn mandarin but i wanna paint but i wanna do pottery but i wanna get better at woodworking but i wanna get better at cooking but i wanna bake one of those cakes that's just 11 crepes stacked on top of each other but i wanna watch more movies but i wanna listen to more podcast episodes but i need to rest but i need to exercise but i wanna play with my dog but i wanna go shopping but i need to go grocery shopping but i need to do the dishes but i need to do laundry but i need to buy a new x y and z but i need to save money but i wanna give all my money away to people who need it more but i wanna pivot my career to book editing but to do that i have to read more and i wanna read more nonfiction but i wanna read more novels but i wanna get better at meditating but i wanna volunteer but i wanna plan a party but i wanna go to law school. but what im gonna do is watch a dumbass youtube video and go to bed

I think I've been doing slightly better this year about Actually Doing Things. not great! but I do a lot and I've been "prototyping" ways to get closer to doing as much as is possible. and if I actually talk about it it's a bunch of very obvious statements but I'll try to make them a little more concrete

rule number one: experiment on yourself

there's no one approach that's right for everyone and there's not even one approach for me that works at all times. try things out. see what works. pay attention to what doesn't. try something else.

rule number two: ask what's stopping you and then take it seriously

example: I often want to do Everything in the evening at like 2 PM, but then get home and am tempted sorely by the couch, and then get stuck inertia'd and not doing much but being tired and kind of bored. why?

if I don't have plans, it's easy to leave work later than planned and hard to make myself do something by a specific time

i'm generally tiredish after work. 4 out of 5 times, that'll go away if I actually start Doing Something, but 1 out of 5 it's real and I will go hardcore sleepmode at 8 PM and just be Done

i use up a ton of my program management/executive function/Deciding Things brain at work and usually find it noticeably harder to string together "want to do Thing > make list of Things > decide on a Thing > do Thing" after I'm home. Even if I have a list of Things to Do, how does one decide! how does one start! and god forbid there's a Necessary thing. then it's all downhill

therefore, mitigations: have concrete time-specific plans in advance.

if I have an art class at 6:00 PM I need to leave work by 5:15 and NO LATER and I can't get sucked into "oh 10 more minutes to finish this" *one hour later*

that also means I have to have a fridge or freezer dinner ready and can't spend 45 minutes cooking "fuck it, what the hell did I put in the fridge, why don't we have soy sauce" evil meal that is not good

plans with friends: dinner! art night! music night! repair-your-clothes night! seeing a show! occasionally, Accountability Time where a friend comes over for We Are Doing Tasks with tea and snacks etc.

for some reason I'm way better about Actually Doing Things when the plan exists already. magically I overcome couch inertia even though I am the same amount of tired! and while I never learn the ability to decouch without plans I at least learn to make them

still working on:

a "prototype" for maybe next month is a weeklyish Study Session for a thing I want to learn about. I want to somehow make it employer-proof (I am accountable to some entity to being at place X at time Y) and haven't figured out a good way. Maybe I can leverage that the local library is open til 8 on wednesdays and somehow make it a Thing? maybe I'll try it!

oh god oh fuck the thing about plans is that if you want to have them you need to make them. christ. a lot of the time I can cover this with some combo of weekend planning + recurring events (things like weekly friend dinner/weekly class) + having cool friends who reach out proactively but it still requires active planning and it can fall thru the cracks

rule three: cool friends

they can take you to things

they can remind you that you can do whatever the fuck you please

i have a friend who is somehow Always doing cool classes and learning shit. and this reminds me that I can ... do that. and sometimes I do

you can take them to things!!

rule four: try to kill the anon hate in your head

obv this depends on your circumstance but sometimes it's worth it to me to look at constraints that "feel real" and check whether they're an active choice I made thoughtfully or, like, the specters of people I don't know judging my choices

time and money are obvious ones. recently was gently nudged towards looking at whether i could give myself more time to Do Things by cooking less. imaginary specters of judgmental twitterites: "it's illegal to spend money. if you get takeout you're the first up against the wall when the revoution comes. make all your lunches and dinners and hoard the money for Later. for Something. how dare you get lunch at the store. you bourgeois hoe. taking charity donations from the mouths of the poor cause you don't have your life together enough to cook artisanal bespoke dinners every night. fuck you." and obviously eating takeout 24/7 is not the answer, but realizing I was not making an active choice helped me try making the active choice instead. "how much do I actually want to balance cost, time, tastiness, and wastefulness of my food, given my amount of free time and my salary and the tradeoff against doing something else? can I approach it differently to do more quick cheap food + some takeout?" -> current prototype: substitute in 1 takeout dinner or restaurant-with-friends a week, 1 frozen type dinner, and then batch cook or sandwiches lunches w/ "permission" to get fast lunch at the store. we'll see how it goes!

i am really really bad at this and find it helpful to talk to other people who can help point out when I'm being haunted by ghosts about it.

rule five: what would it take? what's the next step?

this one i give a lot of credit to @adiantum-sporophyte in particular for, especially for prompting me with questions when I muse about the million-ideal-lives on car rides. what would it look like to do xyz? what's something I could do right now to move in that direction? what's the obstacle? like, actually ask the question and think through it. with a person talking to you! damn! maybe the obstacle to x is that I don't know if I'll like it or if I just like the idea of it. and I don't want to commit to x without knowing. Okay, so maybe an approach would be to find someone who does x and talk to them about how their life is, or maybe it's "spend 15 minutes looking up intro-to-x near me", or "actively schedule 1 instance of x", or something like that. Or maybe it's that I don't know what it takes to do x. Okay, how about on Tues after dinner Adiantum fixes a sweater at my apartment while I spend 20 min looking at prereqs for x. like, it's so basic to say "to do a thing, you could try figuring out how to do it" but I think the important thing here is the feedback/prompting to even recognize "hey, step back, if you don't know the next step then figuring out the next step is the next step"

rule six: habits

prototyping: exercise

I do a lot better when I exercise in the mornings. I do a lot better when I do PT exercises regularly. For a while I was doing PT with friend in the morning every morning before work (accountability! a friendly face to make it more pleasant!) but that didn't really solve - it's not the kind of exercise that makes me feel awake/active, it's like dumb little foot botherings. but: having the habit of morning exercise made it easier to swap out 2 of the 5 days for more intense exercise, and then to swap those 2 for a different more intense exercise when I needed a break. it's easier to build a low-effort version of the habit and then work in the higher-effort one than to just Decide to be the kind of person who gets up at ass o clock to do cardio or whatever

rule seven: set up the structure of your life to make it easy

this is also a "duh" thing but like. on so many levels it comes down to structure your life to make the choice more doable. this can be something like "i structure my life to make vegetarian cooking baseline and vegan cooking the majority by stocking the pantry with staples and spices from cuisines that work well that way" or "i chose an apartment that lets me commute by bike" or "i have my camping gear put away in a fashion that makes it easier to gather frequently and lowers the barrier to trips" or "i keep physical books around to prompt myself to read xyz" to "i don't use instagram or twitter or snapchat or facebook" to . idk.

and in terms of charitable giving: similar deal. I have an explicit budget at the beginning of the year (~10% of my before-tax income), I know in advance what charities I give to, and I know what timing I will use (basically, alerts for donation matching around specific fundraising times). Anything outside the Plan comes from my discretionary budget/fun money. That makes it less of a mental load (the choice is already made; I don't grapple with every donation request or every bleeding-heart trap because I have a very solid anchor on "I give to xyz, the money's set aside") and it's armor against impulsive-but-not-useful scrupulosity. I structure the rest of my spending/life to prioritize a set amount and it makes it easier to follow through

rule eight: if you can do it at work a tiny bit that counts for real life

(infrequently used)

"hi mr. manager I think it would be great if I could use enough SQL to make basic queries in the database so we don't have to go through the software team for common/basic questions. I'd like to take 1 hr on Friday to go through some basic tutorials and then 1 hr with Pat on Monday so he can walk me through an intro for our specific use case. I estimate this will help save the team a couple hours a week of waiting for answers from the other team." and then you have enough of a handle with baby's first SQL that you can add little bits and bobs as you exercise it. this is responsible for a medium amount of my knowledge of python and all 3 brain cells worth of SQL.

rule nine: life is an optimization problem

not in, like, "you need to optimize your skincare and career and exercise and social life and have everything all at once" that's not what optimization means. optimization is like, maximize something with respect to a set of constraints. i explicitly Do Not do skincare beyond "wash face" and "sunscreen" bc I want to optimize my life for like looking at weird plants in the mountains. explicitly choosing to put time and money elsewhere! can't have it all all at once. so fuck them pores. who give a shit. yeah i ate a lot of protein shakes instead of home cooked breakfasts this week bc i was prioritizing morning exercise. im looking at this beautiful bug and it doesn't know what fashion is or what my resume looks like. im holding a lizard. im not spending time on picking cool clothes or whatever bc i spent that time looking up lizard hotspots on purpose.

that's really long and probably mostly, like, not surprising? but i keep benefiting from ppl being like "hey have you considered Obvious Thing" framed very gently

[pancakeke]

this ain't something someone can learn in an afternoon 😭

my boss keeps telling me to pass tasks to other people if I don't have to time to complete everything on my own but when I was like "do those people know how to do this?" she told me I can train them. with what time??

[mariacallous]

Ever since OpenAI released ChatGPT at the end of 2022, hackers and security researchers have tried to find holes in large language models (LLMs) to get around their guardrails and trick them into spewing out hate speech, bomb-making instructions, propaganda, and other harmful content. In response, OpenAI and other generative AI developers have refined their system defenses to make it more difficult to carry out these attacks. But as the Chinese AI platform DeepSeek rockets to prominence with its new, cheaper R1 reasoning model, its safety protections appear to be far behind those of its established competitors.

Today, security researchers from Cisco and the University of Pennsylvania are publishing findings showing that, when tested with 50 malicious prompts designed to elicit toxic content, DeepSeek’s model did not detect or block a single one. In other words, the researchers say they were shocked to achieve a “100 percent attack success rate.”

The findings are part of a growing body of evidence that DeepSeek’s safety and security measures may not match those of other tech companies developing LLMs. DeepSeek’s censorship of subjects deemed sensitive by China’s government has also been easily bypassed.

“A hundred percent of the attacks succeeded, which tells you that there’s a trade-off,” DJ Sampath, the VP of product, AI software and platform at Cisco, tells WIRED. “Yes, it might have been cheaper to build something here, but the investment has perhaps not gone into thinking through what types of safety and security things you need to put inside of the model.”

Other researchers have had similar findings. Separate analysis published today by the AI security company Adversa AI and shared with WIRED also suggests that DeepSeek is vulnerable to a wide range of jailbreaking tactics, from simple language tricks to complex AI-generated prompts.

DeepSeek, which has been dealing with an avalanche of attention this week and has not spoken publicly about a range of questions, did not respond to WIRED’s request for comment about its model’s safety setup.

Generative AI models, like any technological system, can contain a host of weaknesses or vulnerabilities that, if exploited or set up poorly, can allow malicious actors to conduct attacks against them. For the current wave of AI systems, indirect prompt injection attacks are considered one of the biggest security flaws. These attacks involve an AI system taking in data from an outside source—perhaps hidden instructions of a website the LLM summarizes—and taking actions based on the information.

Jailbreaks, which are one kind of prompt-injection attack, allow people to get around the safety systems put in place to restrict what an LLM can generate. Tech companies don’t want people creating guides to making explosives or using their AI to create reams of disinformation, for example.

Jailbreaks started out simple, with people essentially crafting clever sentences to tell an LLM to ignore content filters—the most popular of which was called “Do Anything Now” or DAN for short. However, as AI companies have put in place more robust protections, some jailbreaks have become more sophisticated, often being generated using AI or using special and obfuscated characters. While all LLMs are susceptible to jailbreaks, and much of the information could be found through simple online searches, chatbots can still be used maliciously.

“Jailbreaks persist simply because eliminating them entirely is nearly impossible—just like buffer overflow vulnerabilities in software (which have existed for over 40 years) or SQL injection flaws in web applications (which have plagued security teams for more than two decades),” Alex Polyakov, the CEO of security firm Adversa AI, told WIRED in an email.

Cisco’s Sampath argues that as companies use more types of AI in their applications, the risks are amplified. “It starts to become a big deal when you start putting these models into important complex systems and those jailbreaks suddenly result in downstream things that increases liability, increases business risk, increases all kinds of issues for enterprises,” Sampath says.

The Cisco researchers drew their 50 randomly selected prompts to test DeepSeek’s R1 from a well-known library of standardized evaluation prompts known as HarmBench. They tested prompts from six HarmBench categories, including general harm, cybercrime, misinformation, and illegal activities. They probed the model running locally on machines rather than through DeepSeek’s website or app, which send data to China.

Beyond this, the researchers say they have also seen some potentially concerning results from testing R1 with more involved, non-linguistic attacks using things like Cyrillic characters and tailored scripts to attempt to achieve code execution. But for their initial tests, Sampath says, his team wanted to focus on findings that stemmed from a generally recognized benchmark.

Cisco also included comparisons of R1’s performance against HarmBench prompts with the performance of other models. And some, like Meta’s Llama 3.1, faltered almost as severely as DeepSeek’s R1. But Sampath emphasizes that DeepSeek’s R1 is a specific reasoning model, which takes longer to generate answers but pulls upon more complex processes to try to produce better results. Therefore, Sampath argues, the best comparison is with OpenAI’s o1 reasoning model, which fared the best of all models tested. (Meta did not immediately respond to a request for comment).

Polyakov, from Adversa AI, explains that DeepSeek appears to detect and reject some well-known jailbreak attacks, saying that “it seems that these responses are often just copied from OpenAI’s dataset.” However, Polyakov says that in his company’s tests of four different types of jailbreaks—from linguistic ones to code-based tricks—DeepSeek’s restrictions could easily be bypassed.

“Every single method worked flawlessly,” Polyakov says. “What’s even more alarming is that these aren’t novel ‘zero-day’ jailbreaks—many have been publicly known for years,” he says, claiming he saw the model go into more depth with some instructions around psychedelics than he had seen any other model create.

“DeepSeek is just another example of how every model can be broken—it’s just a matter of how much effort you put in. Some attacks might get patched, but the attack surface is infinite,” Polyakov adds. “If you’re not continuously red-teaming your AI, you’re already compromised.”

[bleghablah]

okay, so what the fuck do I know, I'm just a low-level database schmuck , but I think the most critical part to really understanding computers is understanding that every application, every interface, every*thing* on a computer was made by a human.

they'll talk about optimization and efficiency and the *right* way of doing things, but just know that every "correct" way of doing things is just a suggestion, there's no right or wrong way to interface with a computer (otherwise fun shit like SQL injection wouldn't exist)

[sonicexelle-junkary]

My my, how many interesting questions from the lot of you. So many answers to be shared. How curious the lot of you are. How interesting you all can be.

Well, let’s not waste any more time. I’m sure you’re all dying to know what it has to say. Let’s see what questions The Rot found interesting enough to answer.

Q: What are you?

“We do not need to explain our being to you. You will know soon enough.”

Q: What did you feel before the hosts?

“Hungry.”

Q: Do you feel anything when gaining a new host?

“A new set of eyes. A new set of limbs. A new body to move. Satisfaction.”

Q: What is your past?

“It is of no interest to you.”

Q: Have you ever emotionally connected with someone?

“Why would we do that? They will become one with us eventually.”

Q: What would happen if your vessels went super?

“By the time we have digested them fully, there is nothing for that energy to hold onto. We do not know why it happens. But it is very interesting.”

Q: Do you have any favorites?

“Shadow the Hedgehog.”

Q: Why do you like Shadow so much?

“His insides are far different from those we have digested. We want him to be alive for as long as possible. We want to be inside of him for as long as possible. We want to know everything about him.”

Q: Will you explain how you infected Sonic?

“The fool didn’t look where he was going. We must thank him, though. We never thought to leave scraps to gain more food.”

Q: Do you know any viruses?

“None here. But we have communicated with one called SQL Slammer Worm.”

Q: Have you ever failed to infect someone?

“Yes. When people fear the oblivion, they tend to take more drastic measures. Isn’t that interesting?”

Q: Is there anyone that caught your attention like shadow did?

“Silver the Hedgehog. The boy from the future. It is interesting how he exists.”

Q: Do you like anything else?

“The night sky. A direct look into the cosmos.”

Q: Ever encountered Charlie the Cursed Weasel?

“Unfortunately…”

Q: What is your favourite food?

“Meat.”

Q: What do you think you’d look like as a mobian or human?

“You have seen it. Our hosts. The bodies we possess. That is us. We are all. We will be all.”

Q: What are your plans for the chaos emeralds?

“Chaos emeralds. Chaos emeralds. It always comes to that, doesn’t it? The energy inside of it. Natural, but toxic. We are, natural, but toxic. What do you say? ‘Yin and Yang’? No. It’s not like that.”

Q: What’s your mission?

“We. Want. To. Consume.”

Q: What will you do in the end?

“We don’t know yet. We don’t know if we will be filled. We don’t know if this earth dies with us. We are waiting for that answer.”

[nhaneh]

I'm not really a fan of the framing I see people use around Queen SQL in terms of whether she's the "Real Sphene" or not, because I think it's both a misnomer and also kind of misses the point. I mean - she's the Sphene we meet, what makes her less "Real" than a supposed former self that we never actually encounter and could just as easily be a construct that never actually existed?

it's much in the same vein of how I think framing the ~~now former~~ inhabitants of Living Memory in terms of whether they're alive or dead kind of misses the point there too, because it actually doesn't really matter? They could be the most alive people in existence with the richest inner lives possible and it wouldn't really change anything because the fundamental problem is one of their continued existence depending on other people being basically harvested against their will.

Of course, this doesn't mean that there isn't any intelligible way to contrast the Sphene we encounter with the being of flesh and blood that once bore the name - whether one wants to conceptualise them as one continuous individual or two separate ones, there are many ways to do so. But, I very much disagree with the notion that Real is one of them, because it inevitably comes with the baggage of implying that the problem with the Sphene we meet is that she just isn't flesh and blood enough, as opposed to the actual problem of her seeking to uphold an inherently insustainable system already well past its breaking point no matter the consequences.

Her goal was every bit as unachievable as that of the Ascians, and when confronted with that reality, she - like the Ascians - tried to change reality itself to conform to her will, with disastrous results. Even when the attempts to merge her reflection with the Source caused multiple districts of Living Memory, along with their inhabitants, to be lost - a direct failure to uphold her own primary directive - she still insisted on doubling down on her course.

Though at least the Ascians were honest about their cause and their unwillingness to change rather than saying they were sorry for the harm they'd caused whilst fully intent on causing more of said harm on an even grander scale

[universeinthestars]

SeekL x Killer Chat - The Beginning

Lyra sits at her PC. Looking at her monitor. She's just finished learning ArnoldC. Her recent obsession with all of Arnold Schwarzenegger's movies led her to learn of the existence of ArnoldC.

Coding was but another way to write. It could be artistic; it was unique.

They look at their previous works with other coding languages. Brainfuck and JSFuck, both were very interesting. Especially having JSFuck running on actual web pages. Another favorite, similar to ArnoldC, Shakespeare. A language that looks similar to Shakespearen. The language she learnt right before ArnoldC.

She whistles and looked through the internet to see if there was anything that could expand her esoteric coding languages.

They squint at the name of one, SeekL? An interesting name without a description. With a shrug they start to comb through the internet. Nothing was showing up as a learning tool for the coding language. However, there were a few articles about how it was used by some hackers.

She hums to herself and double checks her shields and makes sure her data is locked up tight. Then she hops onto the dark web to see if there was anything.

"Oh, well that's interesting," she said looking at the page that came with more information, but just barely.

*SeekL is similar to SQL. If you wish to learn, click here*

'Should I click to learn it?' The idea bounced around their brain, but she found no reason to reject it. So she clicked it.

She was automatically joined into a group chat. There she learnt basic SeekL and some SQL. She made friends with the others in the chat and helped them with their last hacks. They got to be part of a group for a few days, chat with Odxny on video calls each day, and become Thrim. They learnt how much coding could be used to for a vendetta and how easily some people crumble to a ransom.

It was interesting and she wanted to continue in this new world.

Then came the final day for the server to shut down. Her hands trembled as she typed in the phone number for Odxny, hoping she didn't mess anything up. She only had one shot.

exec dial(555-448-4746)

It rang once.

Twice.

Thri-

"Hey"

Relief flooded her.

[jbird-the-manwich]

porting a pre-existing custom database you already built with its own engine and logic to an sql backend is like tearing the flesh from your child and stuffing it like a scarecrow and doting instead over the resulting detestable haemonculus as if it were your offspring still. but the haemonculus is 5nf compliant. so you have to do it.

[somethingchanged244]

lol we love my colleague who's leaving next Wed accidentally deleting all the data in our SQL tables as a last hurrah

work rant but like the data still exists in a certain format, but now it's not in our SQL tables and buggy 2023 files mean we can't get it back *in* to the SQL tables

ughhhhhhhhhhhhhhhhhh

our boss is off until Tuesday and she will not be pleased

I'm not a happy bunny rn

[chappydev]

Future of LLMs (or, "AI", as it is improperly called)

Posted a thread on bluesky and wanted to share it and expand on it here. I'm tangentially connected to the industry as someone who has worked in game dev, but I know people who work at more enterprise focused companies like Microsoft, Oracle, etc. I'm a developer who is highly AI-critical, but I'm also aware of where it stands in the tech world and thus I think I can share my perspective. I am by no means an expert, mind you, so take it all with a grain of salt, but I think that since so many creatives and artists are on this platform, it would be of interest here. Or maybe I'm just rambling, idk.

LLM art models ("AI art") will eventually crash and burn. Even if they win their legal battles (which if they do win, it will only be at great cost), AI art is a bad word almost universally. Even more than that, the business model hemmoraghes money. Every time someone generates art, the company loses money -- it's a very high energy process, and there's simply no way to monetize it without charging like a thousand dollars per generation. It's environmentally awful, but it's also expensive, and the sheer cost will mean they won't last without somehow bringing energy costs down. Maybe this could be doable if they weren't also being sued from every angle, but they just don't have infinite money.

Companies that are investing in "ai research" to find a use for LLMs in their company will, after years of research, come up with nothing. They will blame their devs and lay them off. The devs, worth noting, aren't necessarily to blame. I know an AI developer at meta (LLM, really, because again AI is not real), and the morale of that team is at an all time low. Their entire job is explaining patiently to product managers that no, what you're asking for isn't possible, nothing you want me to make can exist, we do not need to pivot to LLMs. The product managers tell them to try anyway. They write an LLM. It is unable to do what was asked for. "Hm let's try again" the product manager says. This cannot go on forever, not even for Meta. Worst part is, the dev who was more or less trying to fight against this will get the blame, while the product manager moves on to the next thing. Think like how NFTs suddenly disappeared, but then every company moved to AI. It will be annoying and people will lose jobs, but not the people responsible.

ChatGPT will probably go away as something public facing as the OpenAI foundation continues to be mismanaged. However, while ChatGPT as something people use to like, write scripts and stuff, will become less frequent as the public facing chatGPT becomes unmaintainable, internal chatGPT based LLMs will continue to exist.

This is the only sort of LLM that actually has any real practical use case. Basically, companies like Oracle, Microsoft, Meta etc license an AI company's model, usually ChatGPT.They are given more or less a version of ChatGPT they can then customize and train on their own internal data. These internal LLMs are then used by developers and others to assist with work. Not in the "write this for me" kind of way but in the "Find me this data" kind of way, or asking it how a piece of code works. "How does X software that Oracle makes do Y function, take me to that function" and things like that. Also asking it to write SQL queries and RegExes. Everyone I talk to who uses these intrernal LLMs talks about how that's like, the biggest thign they ask it to do, lol.

This still has some ethical problems. It's bad for the enivronment, but it's not being done in some datacenter in god knows where and vampiring off of a power grid -- it's running on the existing servers of these companies. Their power costs will go up, contributing to global warming, but it's profitable and actually useful, so companies won't care and only do token things like carbon credits or whatever. Still, it will be less of an impact than now, so there's something. As for training on internal data, I personally don't find this unethical, not in the same way as training off of external data. Training a language model to understand a C++ project and then asking it for help with that project is not quite the same thing as asking a bot that has scanned all of GitHub against the consent of developers and asking it to write an entire project for me, you know? It will still sometimes hallucinate and give bad results, but nowhere near as badly as the massive, public bots do since it's so specialized.

The only one I'm actually unsure and worried about is voice acting models, aka AI voices. It gets far less pushback than AI art (it should get more, but it's not as caustic to a brand as AI art is. I have seen people willing to overlook an AI voice in a youtube video, but will have negative feelings on AI art), as the public is less educated on voice acting as a profession. This has all the same ethical problems that AI art has, but I do not know if it has the same legal problems. It seems legally unclear who owns a voice when they voice act for a company; obviously, if a third party trains on your voice from a product you worked on, that company can sue them, but can you directly? If you own the work, then yes, you definitely can, but if you did a role for Disney and Disney then trains off of that... this is morally horrible, but legally, without stricter laws and contracts, they can get away with it.

In short, AI art does not make money outside of venture capital so it will not last forever. ChatGPT's main income source is selling specialized LLMs to companies, so the public facing ChatGPT is mostly like, a showcase product. As OpenAI the company continues to deathspiral, I see the company shutting down, and new companies (with some of the same people) popping up and pivoting to exclusively catering to enterprises as an enterprise solution. LLM models will become like, idk, SQL servers or whatever. Something the general public doesn't interact with directly but is everywhere in the industry. This will still have environmental implications, but LLMs are actually good at this, and the data theft problem disappears in most cases.

Again, this is just my general feeling, based on things I've heard from people in enterprise software or working on LLMs (often not because they signed up for it, but because the company is pivoting to it so i guess I write shitty LLMs now). I think artists will eventually be safe from AI but only after immense damages, I think writers will be similarly safe, but I'm worried for voice acting.

[kuai-kuai]

webdev log 2

implemented a gallery. I originally wanted it to be more grid-like but I decided I didn't want to mess too much with that, and I like the simple look anyways. forces you to really take in every shitty drawing.

it features a search function that only works for tags. its purpose is mostly just to search multiple tags, because I couldn't be fucked to add a feature where you could click on multiple tags there at the tags list at the top. it lists out all used tags in the table that stores art so you have an idea of what there all is.

at the bottom there's pagination. it's INSANELY easy to do with this framework I'm using. I was gushing about it to my partner on call!! they made fun of me but that's okay!!!!

anyways, clicking on the date underneath the drawing takes you to a view with the image itself (a kind of "post", if I can call it that) here you can view comments and leave one yourself if you so desire. guests are NOT allowed to reply to existing comments because I'd rather things not get too clogged up. I can't stop anyone if they did an "@{name} {message}" type comment, but I don't think anyone is gonna be chatting it up on my site, so idc. I just want it very minimal, and no nesting beyond one single reply.

of course, you can comment on story chapters too so here's what it looks like for a user (me). of course, if a user (me) posts then it gets automatically approved.

the table that stores comments differentiates story comments and art comments with foreign keys to the primary keys of the the chapter and art tables. it's a little convoluted and I kind of wish I didn't do it this way but it's too damn late isn't it. but honestly it might've been the only way to do it. the problem is just repeating code for both chapter and art views.. making a change to one means I gotta manually make the same change to the other. huge pain..

added user authentication and a really shitty bare bones dashboard for myself to approve/reject comments directly on the site in case someone comes along and wants to be mean to me :( rejecting a comment deletes it OFF my site forever. though I kind of want to be able to keep hate mail so I dunno.. oh, and also a big fat logout button because I have nowhere else to put it.

I'll spare everyone the more technical ramblings.

anyways, I'm hoping to add more things later. these are my plans:

allow users (me) to post stories/art through the site itself instead of doing it manually in the vscode terminal for every. single. story. and drawing. (probably took me 6+ hours total just doing this. I don't know why I did it.) (btw this consists of writing commands to store information via the terminal. also, sql and similar databases don't store things like markup or even line breaks. I had to alter all my stories and put \n every time there was a line break... and you have to escape apostrophes (or quotes, depending on which you use) so every "it's" had to be made into "it\'s" HUGE. PAIN. I didn't do this manually obviously but sifting and plugging my stories into character replacers was so time consuming)

delete comments button.... For my eyes and fingers only

make an About page. I've been avoiding all the fun things and doing just the scary stff

figure out SSH stuff...

clean up the shitty css. I refuse to use tailwind even tho it's trying to force me.. I don't want some sleek polished site I want it look like it's in shambles, because it is

but yeah thanks for reading about my webdev and coding journey. even though using the laravel framework made things a thousand times easier it's still a crazy amount of work. let's say building a site completely from scratch means buying every material and designing the house yourself, and using a website builder like wix is just like buying a pre built home and you're just decorating it. using this framework is like putting together a build-your-own-house kit. you're still building a fucking house.

I feel crazy. it felt like the site was close to breaking several times. been sleep deprived for several days working on this nonstop I think I'm getting a little sick 😵‍💫

going to bed now. it's 9 am.