#LLM integration tools

[olivergisttv]

Top Tools for Data Scientists in 2025: What You Must Master to Stay Relevant

The data science profession is facing a massive toolchain shake-up. What worked in 2020 is now considered outdated. With the rise of LLMs, local-first analytics, and AI-native workflows, companies are rethinking their entire data stack. If you’re a data scientist—or hoping to become one—this is your survival guide. Here’s a deep dive into the cutting-edge tools, obsolete tech to avoid, and career…

[chattingwithmodels]

Google Gen AI SDK, Gemini Developer API, and Python 3.13

A Technical Overview and Compatibility Analysis 🧠 TL;DR – Google Gen AI SDK + Gemini API + Python 3.13 Integration 🚀 🔍 Overview Google’s Gen AI SDK and Gemini Developer API provide cutting-edge tools for working with generative AI across text, images, code, audio, and video. The SDK offers a unified interface to interact with Gemini models via both Developer API and Vertex AI 🌐. 🧰 SDK…

[imgblogger]

How to Build a Custom Shopify Page Using AI: A Beginner's Guide to Code-Assisted Development

Ever wanted to create a personalized product page where customers can upload their own images, choose sizes, and preview their creations - all without touching a single line of code? I recently built exactly that using Shopify and AI tools, and I'm here to show you how you can do it too.

Why Use AI for Shopify Development?

Before we dive in, let me share something exciting: AI has revolutionized how we approach web development. Instead of spending months learning to code or hiring expensive developers, you can now use AI as your personal web development assistant. It's like having a senior developer on speed dial, ready to help 24/7.

Understanding Shopify's Theme Editor and Its Limitations

Let's be completely honest about what you can and can't do in Shopify's Theme Editor:

What the Theme Editor Actually Does:

Log into your Shopify account

Click on "Online Store" → "Themes" → "Customize"

Here you can:

Rearrange existing blocks and sections

Change colors and fonts

Adjust spacing and layouts

Add pre-built sections

Modify basic settings

But here's the important part: The Theme Editor alone won't let you build custom features like image uploaders or preview generators. Think of it like decorating a pre-built house - you can change the paint and move the furniture, but you can't add new rooms without some construction work.

The Real Path to Custom Features

To create truly custom features, we need to work with code. Don't panic! Here's how we'll approach this:

Access the Code Editor:

In your theme, click "Actions" → "Edit code"

This is where the real customization happens

You'll see folders and files with different extensions (.liquid, .js, .css)

How to Have a Productive Conversation with AI About Custom Shopify Development

Starting with the Basics

First, let's understand how to break down our conversations with AI into logical steps. Here's the sequence I used to build my custom image upload feature:

Understanding the Tools (First Chat Session)

Act as a Shopify development teacher for complete beginners. I want to create a page where users can upload images and preview them. What tools and access do I need before starting? Please explain each tool's purpose in simple terms.

This will help you understand:

Which Shopify plan you need

Required developer tools

Any third-party services needed

Basic setup requirements

Understanding the Process (New Chat Session)

Act as a Shopify project manager. I want to create an image upload feature. Can you break down this project into small, manageable steps? Please explain what each step accomplishes.

This helps you:

See the big picture

Understand the workflow

Know what to learn first

Plan your development path

Learning About File Structure (New Chat Session)

Act as a Shopify file structure expert. I'm starting from zero. Where in my Shopify theme should I add custom features? Please explain:1. Which folders matter for custom features2. What each important file does3. Where to safely add new code4. What files I should never modify

Getting into Development

Starting the Code (New Chat Session)

Act as a Shopify developer who's great at explaining basics. I need to create an image upload feature. Please help me with:

1. The exact files I need to create 2. Where to put them in my them. 3. Basic structure of each file 4. How these files work together

Building the Upload Feature (New Chat Session)

Act as a Shopify code expert. I'm ready to build the image upload feature. Please provide: 1. The HTML code for the upload button 2. Any required JavaScript 3. Explanation of each code block 4. Where exactly to paste each piece

Creating the Preview Feature (New Chat Session)

Act as a front-end developer. I have the upload feature working. Now I need to: 1. Show a preview of the uploaded image 2. Let users select sizes 3. Display the final previewPlease provide the code and explain how it works.

Important Tips for AI Communication

Be Specific About Your Knowledge Level

Bad example: "How do I add code to Shopify?"

Good example: "I'm new to Shopify and have never modified code before. Please explain how to safely add custom code to my theme, assuming I don't know any technical terms yet."

Ask for Explanations, Not Just Solutions

Bad example: "Give me the code for image upload."

Good example: "Please provide the code for image upload and explain:

What each part does

Why each line is necessary

How to test if it's working

Common mistakes to avoid"

Request Context for Technical Terms

Bad example: "How do I use liquid tags?"

Good example: "I've heard about 'liquid tags' but don't know what they are. Can you:

Explain what liquid tags are in simple terms

Show basic examples

Explain when and why to use them"

Break Down Complex Features When building the image upload feature, ask about each component separately:

First chat: File input creation

New chat: Image preview

New chat: Size selection

New chat: Final preview generation

Real Example of Progressive Questions

Here's how I built my feature, step by step:

First Session:

Act as a Shopify expert for beginners. What exactly happens when a user uploads an image on a Shopify store? Please explain the process in non-technical terms.

Next Session:

Act as a Shopify code teacher. Now that I understand the upload process, how do I create a basic file upload button? Please provide the simplest working example.

Following Session:

Act as a JavaScript expert who's good with beginners. I have a working upload button. How do I show a preview of the uploaded image? Please explain like I'm five.

Testing and Troubleshooting

Always end your AI conversations with:What should I test to make sure this works properly? Please provide: 1. Specific things to check 2. Common problems to watch for 3. How to fix basic issues 4. When to seek more help

Remember: AI is your patient teacher, but you need to ask the right questions. Start with understanding, then move to implementation, and always ask for explanations rather than just code.

Success Stories

When I first started, I didn't know anything about Shopify or coding. Using these exact steps, I created a custom page where customers can:

Upload their pictures

Choose sticker sizes

Select quantities

See exactly how their final product will look

[ai-network]

LangChain: Components, Benefits & Getting Started

Understanding the Core Components of LangChain

LangChain is a revolutionary framework designed to enhance the capabilities of Large Language Models (LLMs) by enabling them to process and comprehend real-time data more efficiently. At its core, LangChain is built on foundational components that support its robust architecture. These components include: - Data Connectors: These facilitate seamless integration with various data sources, allowing LLMs to access diverse datasets in real-time. - Processing Pipelines: LangChain employs sophisticated pipelines that preprocess and transform raw data into structured formats suitable for consumption by LLMs. - Semantic Parsers: These components help interpret and extract meaningful information from text inputs, providing LLMs with context-rich data. - Inference Engines: At the heart of LangChain, inference engines leverage advanced algorithms to derive insights from the processed data, enhancing the decision-making capabilities of LLMs. Together, these components form an integrated ecosystem that empowers developers to build dynamic, AI-driven applications.

How LangChain Enhances LLM Capabilities with Real-Time Data

One of the standout features of this framework is its ability to augment LLM capabilities through real-time data integration. Traditional language models often operate in static environments, relying on pre-trained data sets. However, LangChain breaks this limitation by establishing live connections with dynamic data sources. Using its advanced data connectors, it can pull data from APIs, databases, and streams, ensuring that LLMs are informed by the most current information available. This real-time data ingestion not only increases the relevancy of LLM outputs but also enables adaptive learning. The synchronous feeding of real-time data into LLMs allows applications powered by LangChain to react swiftly to changes, whether they pertain to market trends, news events, or user interactions. By leveraging real-time data, LangChain truly sets itself apart as a tool for modern AI applications, providing both accuracy and agility in decision-making processes.

Streamlining Data Organization for Efficient LLM Access

Efficiency in accessing and processing data is crucial for optimizing the performance of LLMs. LangChain introduces several methodologies to streamline data organization, thereby facilitating quick and efficient data retrieval. Firstly, the framework implements a hierarchical data storage system that categorizes data based on its relevance and frequency of access. This enables the prioritization of data that is most pertinent to ongoing tasks, reducing latency in information retrieval. Secondly, LangChain employs advanced indexing techniques. By creating indices tailored to specific data attributes, LangChain accelerates the search process, enabling LLMs to access necessary data rapidly. Furthermore, the use of semantic tagging enhances this process, allowing for intelligent filtering based on contextually relevant keywords. Lastly, a commitment to data normalization within LangChain ensures that data from disparate sources is harmonized into a uniform format. This standardization minimizes the complexity during data processing stages and allows LLMs to interpret data consistently, leading to more accurate results.

Step-by-Step Guide to Developing LLM-Powered Applications with LangChain

Developing applications powered by LangChain involves a systematic approach that maximizes the potential of LLMs. Here is a step-by-step guide to help developers get started: - Define Application Objectives: Clearly outline the goals of your application, particularly how it will utilize LLMs to achieve these objectives. - Select Appropriate Data Sources: Choose data sources that align with your application’s objectives. LangChain’s data connectors support a wide range of sources, including APIs and databases. - Configure Data Connectors: Set up the data connectors in LangChain to establish live feeds from your chosen data sources, ensuring real-time data availability. - Design the Processing Pipeline: Construct a data processing pipeline within LangChain to handle data transformations and preprocessing requirements specific to your application. - Implement Semantic Parsing: Integrate semantic parsers to enrich your data with contextual meaning and facilitate comprehensive interpretation by the LLMs. - Develop Inference Mechanisms: Build inference mechanisms using LangChain’s inference engines to derive actionable insights from the processed data. - Prototype and Test: Develop a prototype of your application and conduct thorough testing to validate functionality and ensure reliability. - Iterate and Optimize: Continuously iterate on your design, incorporating feedback and optimizing components for improved performance. This structured approach not only streamlines the development process but also ensures that the resulting application harnesses the power of LangChain efficiently.

Maximizing the Potential of LangChain in Modern AI Development

In today’s rapidly evolving technological landscape, the potential of LangChain in modern AI development is immense. Its unique combination of real-time data integration, robust processing capabilities, and compatibility with large language models position it as an indispensable tool for developers. To maximize its potential, developers should focus on tailoring LangChain's capabilities to their specific use cases. By aligning LangChain’s powerful functionalities with the unique requirements of their applications, developers can create highly specialized AI solutions that deliver exceptional value. Additionally, staying abreast of updates and enhancements to LangChain will ensure that developers leverage the latest features and improvements. Engaging with the LangChain community, participating in forums, and accessing documentation can provide valuable insights and support. Finally, experimentation and innovation are key. By exploring novel approaches and pushing the boundaries of what is possible with LangChain, developers can unlock new levels of sophistication in AI-driven applications, driving forward the future of AI technology. In conclusion, LangChain stands out as a transformative framework in AI development, offering a suite of tools and components that empower developers to build intelligent, responsive applications. By understanding and implementing its capabilities strategically, one can fully harness its potential to drive innovation in the field of artificial intelligence. Read the full article

[ellipsus-writes]

Back when we started Ellipsus (it's been eighty-four years… or two, but it sure feels like forever), we encountered generative AI.

Immediately, we realized LLMs were the antithesis of creativity and community, and the threat they posed to genuine artistic expression and collaboration. (P.S.: we have a lot to say about it.)

Since then, writing tools—from big tech entities like Google Docs and Microsoft Word, to a host of smaller platforms and publishers—have rapidly integrated LLMs, looking to capitalize on the novelty of generative AI. Now, our tools are failing us, corrupted by data-scraping and hostile to users' consent and IP ownership.

The future of creative work requires a nuanced understanding of the challenges ahead, and a shared vision—writers for writers. We know we're stronger together. And in a rapidly changing world, we know that transparency is paramount.

So… some Ellipsus facts:

We will never include generative AI in Ellipsus.

We will never access your work without explicit consent, sell your data, or use your work for exploitative purposes.

We believe in the strength of creative communities and the stories they tell—and we want to foster a space in which writers can connect and tell their stories in freedom and safety—without compromise.

[vlruso]

Unlocking Multimodal AI with Open AI: GPT-4Vs Vision Integration and Its Impact

📢 Introducing Open AI's GPT-4V: the next level of AI! 🌟 With vision integration, this language model analyzes images alongside text, unlocking new opportunities and challenges. 🌐 Find out how it works and its impact here: https://ift.tt/Wvotjzu. #AI #GPT4V #OpenAI List of Useful Links: AI Scrum Bot - ask about AI scrum and agile Our Telegram @itinai Twitter -  @itinaicom

[fandomtrumpshate]

Our Stance On Gen-AI

This year, for the first time, we've had a couple of reports from bidders that the FTH fanworks they received were produced using generative AI. For that reason, we've decided that it's important that we lay out a specific, concrete policy going forward.

Generative AI tools are not welcome here.

Non-exhaustive list of examples:

image generators like Imagen, Midjourney, and similar

video generators like Sora, Runway, and similar

LLMs like ChatGPT and similar

audio generators like ElevenLabs, MusicLM, and similar

Participants found to have used generative AI to produce a fanwork, in part or in whole, for their bidder(s) will be permanently banned from participating in future iterations of Fandom Trumps Hate.

Why?

We understand that there can be contentious debate around the use of generative AI, we know individual people have their own reasons for being in favor of it, and we recognize that many people may simply be unaware that these tools come with any negative impacts at all. Regardless, we are firm in our stance on this for the following (non-exhaustive) list of key reasons in no particular order:

negative, unregulated environmental impact

Over the years, you may have noticed that we’ve supported multiple environmental organizations doing important work to combat climate change, preserve wildlife, and advocate for renewable and sustainable energy policy changes. Generative AI tools produce a startling amount of e-waste, can require massive amounts of storage space and computational power, and are a (currently unregulated) drain on natural resources. Using these tools to produce a fanwork flies in the face of every environmental organization we have supported to date.

plagiarism and lack of artistic integrity

Most if not all generative AI models are trained on some amount of stolen work (across various mediums). As a result, any output generated by these models is at worst plagiarized and at best extremely derivative and unoriginal. In our opinion, using generative AI tools to produce a fanwork demonstrates a lack of care for your own craft, a lack of respect for the work of other creators, and a lack of respect for your bidder and your commitment to them.

undermining our community building impact

One of the best things to come out of the auction every year—we can't even call it a side benefit, because it's so central to us—is that bidders and creators form collaborative relationships which sometimes even turn into friendship. Using generative AI undermines that trust and collaboration.

undermining the value of participating as a creator

Bidders participate in Fandom Trumps Hate for the opportunity to prompt YOU to create a fanwork for them, in YOUR style with YOUR specific skill set. Any potential bidder is perfectly capable of dropping a prompt into a generative AI tool on their own time, if they wish. We hope all creators sign up with the aim to play a role more significant than “unnecessary middleman.”

In general, we try to be as flexible as we can in our policies to allow for the best experience possible for all Fandom Trumps Hate participants. This, however, is something we are not willing to be flexible on. We realize this may seem unusually rigid, but we ask that you trust we have given this serious consideration and respect that while we are willing to answer clarifying questions, we are not open to debate on this topic.

[imsobadatnicknames2]

How can you consider yourself any sort of leftist when you defend AI art bullshit? You literally simp for AI techbros and have the gall to pretend you're against big corporations?? Get fucked

I don't "defend" AI art. I think a particular old post of mine that a lot of people tend to read in bad faith must be making the rounds again lmao.

Took me a good while to reply to this because you know what? I decided to make something positive out of this and use this as an opportunity to outline what I ACTUALLY believe about AI art. If anyone seeing this decides to read it in good or bad faith... Welp, your choice I guess.

I have several criticisms of the way the proliferation of AI art generators and LLMs is making a lot of things worse. Some of these are things I have voiced in the past, some of these are things I haven't until now:

Most image and text AI generators are fine-tuned to produce nothing but the most agreeable, generically pretty content slop, pretty much immediately squandering their potential to be used as genuinely interesting artistic tools with anything to offer in terms of a unique aesthetic experience (AI video still manages to look bizarre and interesting but it's getting there too)

In the entertainment industry and a lot of other fields, AI image generation is getting incorporated into production pipelines in ways that lead to the immiseration of working artists, being used to justify either lower wages or straight-up layoffs, and this is something that needs to be fought against. That's why I unconditionally supported the SAG-AFTRA strikes last year and will unconditionally support any collective action to address AI art as a concrete labor issue

In most fields where it's being integrated, AI art is vastly inferior to human artists in any use case where you need anything other than to make a superficially pretty picture really fast. If you need to do anything like ask for revisions or minor corrections, give very specific descriptions of how objects and people are interacting with each other, or just like. generate several pictures of the same thing and have them stay consistent with each other, you NEED human artists and it's preposterous to think they can be replaced by AI.

There is a lot of art on the internet that consists of the most generically pretty, cookie-cutter anime waifu-adjacent slop that has zero artistic or emotional value to either the people seeing it or the person churning it out, and while this certainly was A Thing before the advent of AI art generators, generative AI has made it extremely easy to become the kind of person who churns it out and floods online art spaces with it.

Similarly, LLMs make it extremely easy to generate massive volumes of texts, pages, articles, listicles and what have you that are generic vapid SEO-friendly pap at best and bizzarre nonsense misinformation at worst, drowning useful information in a sea of vapid noise and rendering internet searches increasingly useless.

The way LLMs are being incorporated into customer service and similar services not only, again, encourages further immiseration of customer service workers, but it's also completely useless for most customers.

A very annoyingly vocal part the population of AI art enthusiasts, fanatics and promoters do tend to talk about it in a way that directly or indirectly demeans the merit and skill of human artists and implies that they think of anyone who sees anything worthwile in the process of creation itself rather than the end product as stupid or deluded.

So you can probably tell by now that I don't hold AI art or writing in very high regard. However (and here's the part that'll get me called an AI techbro, or get people telling me that I'm just jealous of REAL artists because I lack the drive to create art of my own, or whatever else) I do have some criticisms of the way people have been responding to it, and have voiced such criticisms in the past.

I think a lot of the opposition to AI art has critstallized around unexamined gut reactions, whipping up a moral panic, and pressure to outwardly display an acceptable level of disdain for it. And in particular I think this climate has made a lot of people very prone to either uncritically entertain and adopt regressive ideas about Intellectual Propety, OR reveal previously held regressive ideas about Intellectual Property that are now suddenly more socially acceptable to express:

(I wanna preface this section by stating that I'm a staunch intellectual property abolitionist for the same reason I'm a private property abolitionist. If you think the existence of intellectual property is a good thing, a lot of my ideas about a lot of stuff are gonna be unpalatable to you. Not much I can do about it.)

A lot of people are suddenly throwing their support behind any proposal that promises stricter copyright regulations to combat AI art, when a lot of these also have the potential to severely udnermine fair use laws and fuck over a lot of independent artist for the benefit of big companies.

It was very worrying to see a lot of fanfic authors in particular clap for the George R R Martin OpenAI lawsuit because well... a lot of them don't realize that fanfic is a hobby that's in a position that's VERY legally precarious at best, that legally speaking using someone else's characters in your fanfic is as much of a violation of copyright law as straight up stealing entire passages, and that any regulation that can be used against the latter can be extended against the former.

Similarly, a lot of artists were cheering for the lawsuit against AI art models trained to mimic the style of specific artists. Which I agree is an extremely scummy thing to do (just like a human artist making a living from ripping off someone else's work is also extremely scummy), but I don't think every scummy act necessarily needs to be punishable by law, and some of them would in fact leave people worse off if they were. All this to say: If you are an artist, and ESPECIALLY a fan artist, trust me. You DON'T wanna live in a world where there's precedent for people's artstyles to be considered intellectual property in any legally enforceable way. I know you wanna hurt AI art people but this is one avenue that's not worth it.

Especially worrying to me as an indie musician has been to see people mention the strict copyright laws of the music industry as a positive thing that they wanna emulate. "this would never happen in the music industry because they value their artists copyright" idk maybe this is a the grass is greener type of situation but I'm telling you, you DON'T wanna live in a world where copyright law in the visual arts world works the way it does in the music industry. It's not worth it.

I've seen at least one person compare AI art model training to music sampling and say "there's a reason why they cracked down on sampling" as if the death of sampling due to stricter copyright laws was a good thing and not literally one of the worst things to happen in the history of music which nearly destroyed several primarily black music genres. Of course this is anecdotal because it's just One Guy I Saw Once, but you can see what I mean about how uncritical support for copyright law as a tool against AI can lead people to adopt increasingly regressive ideas about copyright.

Similarly, I've seen at least one person go "you know what? Collages should be considered art theft too, fuck you" over an argument where someone else compared AI art to collages. Again, same point as above.

Similarly, I take issue with the way a lot of people seem EXTREMELY personally invested in proving AI art is Not Real Art. I not only find this discussion unproductive, but also similarly dangerously prone to validating very reactionary ideas about The Nature Of Art that shouldn't really be entertained. Also it's a discussion rife with intellectual dishonesty and unevenly applied definition and standards.

When a lot of people present the argument of AI art not being art because the definition of art is this and that, they try to pretend that this is the definition of art the've always operated under and believed in, even when a lot of the time it's blatantly obvious that they're constructing their definition on the spot and deliberately trying to do so in such a way that it doesn't include AI art.

They never succeed at it, btw. I've seen several dozen different "AI art isn't art because art is [definition]". I've seen exactly zero of those where trying to seriously apply that definition in any context outside of trying to prove AI art isn't art doesn't end up in it accidentally excluding one or more non-AI artforms, usually reflecting the author's blindspots with regard to the different forms of artistic expression.

(However, this is moot because, again, these are rarely definitions that these people actually believe in or adhere to outside of trying to win "Is AI art real art?" discussions.)

Especially worrying when the definition they construct is built around stuff like Effort or Skill or Dedication or The Divine Human Spirit. You would not be happy about the kinds of art that have traditionally been excluded from Real Art using similar definitions.

Seriously when everyone was celebrating that the Catholic Church came out to say AI art isn't real art and sharing it as if it was validating and not Extremely Worrying that the arguments they'd been using against AI art sounded nearly identical to things TradCaths believe I was like. Well alright :T You can make all the "I never thought I'd die fighting side by side with a catholic" legolas and gimli memes you want, but it won't change the fact that the argument being made by the catholic church was a profoundly conservative one and nearly identical to arguments used to dismiss the artistic merit of certain forms of "degenerate" art and everyone was just uncritically sharing it, completely unconcerned with what kind of worldview they were lending validity to by sharing it.

Remember when the discourse about the Gay Sex cats pic was going on? One of the things I remember the most from that time was when someone went "Tell me a definition of art that excludes this picture without also excluding Fountain by Duchamp" and how just. Literally no one was able to do it. A LOT of people tried to argue some variation of "Well, Fountain is art and this image isn't because what turns fountain into art is Intent. Duchamp's choice to show a urinal at an art gallery as if it was art confers it an element of artistic intent that this image lacks" when like. Didn't by that same logic OP's choice to post the image on tumblr as if it was art also confer it artistic intent in the same way? Didn't that argument actually kinda end up accidentally validating the artistic status of every piece of AI art ever posted on social media? That moment it clicked for me that a lot of these definitions require applying certain concepts extremely selectively in order to make sense for the people using them.

A lot of people also try to argue it isn't Real Art based on the fact that most AI art is vapid but like. If being vapid definitionally excludes something from being art you're going to have to exclude a whooole lot of stuff along with it. AI art is vapid. A lot of art is too, I don't think this argument works either.

Like, look, I'm not really invested in trying to argue in favor of The Artistic Merits of AI art but I also find it extremely hard to ignore how trying to categorically define AI art as Not Real Art not only is unproductive but also requires either a) applying certain parts of your definition of art extremely selectively, b) constructing a definition of art so convoluted and full of weird caveats as to be functionally useless, or c) validating extremely reactionary conservative ideas about what Real Art is.

Some stray thoughts that don't fit any of the above sections.

I've occassionally seen people respond to AI art being used for shitposts like "A lot of people have affordable commissions, you could have paid someone like $30 to draw this for you instead of using the plagiarism algorithm and exploiting the work of real artists" and sorry but if you consider paying an artist a rate that amounts to like $5 for several hours of work a LESS exploitative alternative I think you've got something fucked up going on with your priorities.

Also it's kinda funny when people comment on the aforementioned shitposts with some variation of "see, the usage of AI art robs it of all humor because the thing that makes shitposts funny is when you consider the fact that someone would spend so much time and effort in something so stupid" because like. Yeah that is part of the humor SOMETIMES but also people share and laugh at low effort shitposts all the time. Again you're constructing a definition that you don't actually believe in anywhere outside of this type of conversations. Just say you don't like that it's AI art because you think it's morally wrong and stop being disingenuous.

So yeah, this is pretty much everything I believe about the topic.

I don't "defend" AI art, but my opposition to it is firmly rooted in my principles, and that means I refuse to uncritically accept any anti-AI art argument that goes against those same principles.

If you think not accepting and parroting every Anti-AI art argument I encounter because some of them are ideologically rooted in things I disagree with makes me indistinguishable from "AI techbros" you're working under a fucked up dichotomy.

[morlock-holmes]

Man, the AI conversation is so fucking weird.

Tech companies have shoved LLMs and image generators into every nook and cranny they can find, not to mention the fact that there are freeware ones as well.

Anybody reading this can play with AI themselves to find out what it can and can't do, it's right at your fingertips.

And pretty much the entire public conversation about the technology is just totally divorced from the product as it exists.

Instead, various interest groups are waging intense battles over what to do with products that exist entirely in their imagination.

Another review youtuber I otherwise like is doing the "AI is disgusting! Can you imagine somebody whose business model involved using copyrighted content to produce derivative works without permission and then distributing that work through massive data centers?" with just... seemingly no sense of irony whatsoever. Meanwhile, self-same youtuber directed me to Etsy's frankly bizarre AI rules:

While we allow the use of AI tools in the creative process, we prohibit the sale of AI prompt bundles on our platform. We believe that the prompts used to generate AI artwork are an integral part of the creative process and should not be sold separately from the final artwork. Selling prompt bundles without the accompanying finished artwork undermines the value of the artist's creative input and curation, which are essential to the creation of the unique, creative items that Etsy is known for.

Which, like... Okay...

Look, I'm no PR expert, but what was stopping them from saying,

"Given the incredibly widespread use of AI and the proliferation of models which all respond differently, prompt bundles are likely to have extremely limited value, and help with prompting can be found for free at numerous places."

I mean here's the thing: Prompt bundles only have value if AI technology doesn't improve. Like I don't think they ought to be cluttering up Etsy, there's enough garbage on there as it is and there are plenty of other places to find or purchase that stuff if you want it.

Hell you can ask an AI to come up with some prompts for you on a given subject.

I don't know, it's so disorienting that the entire conversation around something so directly accessible is so entirely divorced from the actual, tangible thing, and instead waged around what it might be in five years or what it was two years ago.

It's right here! We have it now! Why has that had so little effect on how we think about it?

[jstor]

I saw something about generative AI on JSTOR. Can you confirm whether you really are implementing it and explain why? I’m pretty sure most of your userbase hates AI.

A generative AI/machine learning research tool on JSTOR is currently in beta, meaning that it's not fully integrated into the platform. This is an opportunity to determine how this technology may be helpful in parsing through dense academic texts to make them more accessible and gauge their relevancy.

To JSTOR, this is primarily a learning experience. We're looking at how beta users are engaging with the tool and the results that the tool is producing to get a sense of its place in academia.

In order to understand what we're doing a bit more, it may help to take a look at what the tool actually does. From a recent blog post:

Content evaluation

Problem: Traditionally, researchers rely on metadata, abstracts, and the first few pages of an article to evaluate its relevance to their work. In humanities and social sciences scholarship, which makes up the majority of JSTOR’s content, many items lack abstracts, meaning scholars in these areas (who in turn are our core cohort of users) have one less option for efficient evaluation. 

When using a traditional keyword search in a scholarly database, a query might return thousands of articles that a user needs significant time and considerable skill to wade through, simply to ascertain which might in fact be relevant to what they’re looking for, before beginning their search in earnest.

Solution: We’ve introduced two capabilities to help make evaluation more efficient, with the aim of opening the researcher’s time for deeper reading and analysis:

Summarize, which appears in the tool interface as “What is this text about,” provides users with concise descriptions of key document points. On the back-end, we’ve optimized the Large Language Model (LLM) prompt for a concise but thorough response, taking on the task of prompt engineering for the user by providing advanced direction to:

Extract the background, purpose, and motivations of the text provided.

Capture the intent of the author without drawing conclusions.

Limit the response to a short paragraph to provide the most important ideas presented in the text.

Search term context is automatically generated as soon as a user opens a text from search results, and provides information on how that text relates to the search terms the user has used. Whereas the summary allows the user to quickly assess what the item is about, this feature takes evaluation to the next level by automatically telling the user how the item is related to their search query, streamlining the evaluation process.

Discovering new paths for exploration

Problem: Once a researcher has discovered content of value to their work, it’s not always easy to know where to go from there. While JSTOR provides some resources, including a “Cited by” list as well as related texts and images, these pathways are limited in scope and not available for all texts. Especially for novice researchers, or those just getting started on a new project or exploring a novel area of literature, it can be needlessly difficult and frustrating to gain traction. 

Solution: Two capabilities make further exploration less cumbersome, paving a smoother path for researchers to follow a line of inquiry:

Recommended topics are designed to assist users, particularly those who may be less familiar with certain concepts, by helping them identify additional search terms or refine and narrow their existing searches. This feature generates a list of up to 10 potential related search queries based on the document’s content. Researchers can simply click to run these searches.

Related content empowers users in two significant ways. First, it aids in quickly assessing the relevance of the current item by presenting a list of up to 10 conceptually similar items on JSTOR. This allows users to gauge the document’s helpfulness based on its relation to other relevant content. Second, this feature provides a pathway to more content, especially materials that may not have surfaced in the initial search. By generating a list of related items, complete with metadata and direct links, users can extend their research journey, uncovering additional sources that align with their interests and questions.

Supporting comprehension

Problem: You think you have found something that could be helpful for your work. It’s time to settle in and read the full document… working through the details, making sure they make sense, figuring out how they fit into your thesis, etc. This all takes time and can be tedious, especially when working through many items. 

Solution: To help ensure that users find high quality items, the tool incorporates a conversational element that allows users to query specific points of interest. This functionality, reminiscent of CTRL+F but for concepts, offers a quicker alternative to reading through lengthy documents. 

By asking questions that can be answered by the text, users receive responses only if the information is present. The conversational interface adds an accessibility layer as well, making the tool more user-friendly and tailored to the diverse needs of the JSTOR user community.

Credibility and source transparency

We knew that, for an AI-powered tool to truly address user problems, it would need to be held to extremely high standards of credibility and transparency. On the credibility side, JSTOR’s AI tool uses only the content of the item being viewed to generate answers to questions, effectively reducing hallucinations and misinformation. 

On the transparency front, responses include inline references that highlight the specific snippet of text used, along with a link to the source page. This makes it clear to the user where the response came from (and that it is a credible source) and also helps them find the most relevant parts of the text. 

[olivergisttv]

Data Science in 2025: The Shocking Tools That Will Make or Break Your Career

The data science field in 2025 is no longer what it used to be. With AI-native tooling, vector search pipelines, and local-first compute on the rise, a seismic shift is underway. 📉 40% of traditional data science tasks are now automated 📈 Roles demand real-time, AI-integrated, and app-ready solutions ❗ And if you’re still clinging to legacy tools like Pandas and Jupyter? You’re not just…

[kunosoura]

You aren’t putting the cat back in the bag on LLMs and TTIMs, sorry. You could have the People’s Army For The Protection Of Intellectual Property Rights And Class Signifiers Artists And Academic Integrity seize and destroy the servers for every major player in “AI” overnight and within a week people would be circulating and hacking the most recent offline models to fill the same niches. You might put a dent in the sheer volume of its usage but there are models that run on consumer laptops now. “AI” art tools, from full image generation to photo editing, are here to stay. College essays and internet content are definitely gonna keep being written by LLM and with tools that help spoof a personal writing style, it will always be a losing arms race to try and reliably catch it. Find a better way to address the underlying problems or kick rocks.

[dragonnarrative-writes]

Is there any like studies about ai stealing work from creators? Like obviously it's a complete violation to put someone elses work into ai, but tbh its a bit helpful sometimes? Like if im studying and theres no answer key ai can integrate pretty well and give steps explaining how to solve ? Like im trying to understand the completely anti ai pov a bit more, bc im not gonna stop using a useful tool wo evidence of wrongdoing

This ask is fascinating to me. Are there any studies about the basic foundations of how generative AI works? Yeah...? If you're already using AI tools, you could even ask the AI and it would tell you. But... OpenAI already said that LLMs are trained on copyrighted works.

If you're studying, and there's no answer key, then you can use already existing tools to help you learn how to conceptualize the problem. Khan Academy has so many resources, and if you don't like those, there are probably a lot of resources available via your library, in person and online. That's not even touching on the fact that you could probably work with your instructor or school to get some help.

I thought about saying a lot more, but even if I assume you asked this in good faith, I don't really think you care. You've already decided that a "complete violation" is "a bit helpful sometimes." You're telling me you see using AI as a convenient and fast tool, and nothing is going to fit the bill more than the Plagiarism Machine.

[reimenaashelyee]

Hello what do you think of Ai generated artwork and videos?

I have a whole entire blog post I wrote last year btw: The Rise of the Bots; The Ascension of the Human. (Reading it again a year later I am glad I am still validated in my thoughts)

My entire being and output as an artist is rooted in process, thought, craft and connection. I am open about my process and I share/create resources constantly. I have literally experienced the thing people mean when they say 'art transforms you' just by being so close to it every step of its making. All my comics have this centrality of personhood attached to them - if it's not obvious that the artist's hand (me) is in it, there is the characteristic focus on our emotional/cultural/artistic thread across history. Just as NFTs and what they represent were antithetical to how I interact with the world as artist and audience, so is the use of so-called AI art. NFTs and AI Art share a common hype cycle / speculative mania that comes out from an annoying vulture mindset that only knows how to eat itself to fill its belly, so I don't expect it to last too long. However I don't appreciate the damage both things have done to the utility of the internet, the degradation of art as a commercial pathway and the destruction of the image as a historical/educational/legal tool. (Which is why I am becoming more underground and turning towards alternatives like the Web Revival, small presses, curated resources and in-person communities)

The technological concept around LLM (pattern recognition and matching it to a goal), especially for medicine and statistics, is not itself problematic, especially when it follows ethical and data handling regulations that have been defined. However, when people talk generative art, what we are talking about, and fighting against, is the exploitation of resources and labour, and the further disconnection of worker = labour, human = society artificially imposed by the Corporate MBA / techno class in the pursuit of infinite stockmarket growth which then introduces a type of brainrot that can only think of things as producing value in relation to how fast one can seize for themselves Westernised Ideals of Fame and Fortune. Also like, this whole AI thing is part of the degradation of entertainment (the loss of small-to-medium outlets, constant mergers, nobody owning their digital streaming products they bought, the laundering of journalism/curation into press releases), the internet (the algorithimification of everything, constant spam, search engines getting worse, the worsening of socmedia as a tool) and the intellectual rigour of all information.

It's all part of this rot that's spreading outwards.

TL;DR bro I make all my art by hand and I am a nerd about informational integrity

[noticiassincensura]

Former OpenAI Researcher Accuses the Company of Copyright Law Violations

Use of Copyrighted Data in AI Models In a new twist in the world of artificial intelligence, Suchir Balaji, a former researcher at OpenAI, has spoken publicly about the company’s practices and its use of copyrighted data. Balaji, who spent nearly four years working at OpenAI, helped collect and organize large volumes of internet data to train AI models like ChatGPT. However, after reflecting on the legal and ethical implications of this process, he decided to leave the company in August 2024.

What Motivated His Departure? Balaji, 25, admitted that at first, he did not question whether OpenAI had the legal right to use the data it was collecting, much of which was protected by copyright. He assumed that since it was publicly available information on the internet, it was free to use. However, over time, and especially after the launch of ChatGPT in 2022, he began to doubt the legality and ethics of these practices.

“If you believe what I believe, you have to leave the company,” he commented in a series of interviews with The New York Times. For Balaji, using copyrighted data without the creators’ consent was not only a violation of the law but also a threat to the integrity of the internet. This realization led him to resign, although he has not taken another job yet and is currently working on personal projects.

A Growing Problem in AI Concerns about the use of protected data to train AI models are not new. Since companies like OpenAI and other startups began launching tools based on large language models (LLMs), legal and ethical issues have been at the forefront of the debate. These models are trained using vast amounts of text from the internet, often without respecting copyright or seeking the consent of the original content creators.

Balaji is not the only one to raise his voice on this matter. A former vice president of Stability AI, a startup specializing in generative image and audio technologies, has also expressed similar concerns, arguing that using data without authorization is harmful to the industry and society as a whole.

The Impact on the Future of AI Such criticisms raise questions about the future of artificial intelligence and its relationship with copyright laws. As AI models continue to evolve, the pressure on companies to develop ethical and legal technologies is increasing. The case of Balaji and other experts who have decided to step down signals that the AI industry might be facing a significant shift in how it approaches data usage.

The conversation about copyright in AI is far from over, and it seems that this will be a central topic in future discussions about the regulation and development of generative technologies

[maxsmith007-blog]

What is Vanie LLM and how is it different from traditional AI models for contact centers?

Contact centers today face significant demands to reduce costs, enhance operations, and deliver exceptional customer experiences. Many have turned to AI, but most traditional systems rely on fixed rules or narrow machine learning that cannot keep up with the complexity of real conversations.

Vanie LLM changes this. It’s a large language model designed specifically for contact centers, offering a smarter, more flexible approach than older AI models. Unlike standard solutions that follow rigid scripts. Vanie LLM understands full conversations and adapts in real time.

How Vanie LLM stands out and impacts business

1. Tracks the entire conversation, not just keywords

Most traditional AI tools only look at short snippets or identify keywords. They miss the full back-and-forth flow of a conversation. That’s a major gap, as 65% of AI-driven responses from legacy systems fail to solve complex, multi-turn issues (Source: CCW Market Study 2025).

Vanie LLM follows the entire exchange, keeping track of context across multiple questions and answers. This helps it understand customer needs better and boosts first-contact resolution.

2. Learns from unstructured data automatically

Older AI models often depend on clean, structured data and predefined categories. They struggle to adjust without expensive manual retraining.

Vanie LLM learns directly from unstructured data—like call transcripts, chat logs, and CRM notes. This means it updates its knowledge quickly and at a lower cost.

Businesses using models like Vanie LLM have seen:

37% improvement in average handle time (AHT)

Fewer call escalations

Better operational KPIs overall (Forrester, 2025)

3. Detects subtle sentiment and compliance risks

Traditional AI may tag basic sentiment—positive or negative—but misses more subtle signals. Vanie LLM can detect complex emotions, signs of dissatisfaction, and compliance issues right as they happen.

This leads to fewer regulatory problems. For example, a contact center serving regulated industries reduced compliance breaches by 28% after using real-time monitoring with advanced language models.

4. Scales easily across all channels

Legacy AI often needs different setups for each channel—voice, chat, or email—creating silos and higher maintenance costs.

Vanie LLM works seamlessly across voice calls, chat, email and even social messaging. It helps unify the tech stack and lowers integration overhead.

Operational gains and clear ROI

Switching to LLM-driven systems delivers direct, measurable results:

15-25% lower operational costs from faster calls, less manual QA, and reduced workload

20%+ increase in CSAT scores with quicker, more accurate, and more personal service

30% faster agent ramp-up times, thanks to intelligent guidance that supports new hires from day one

Vanie LLM also gives leaders powerful insights into agent performance, common customer issues, and process bottlenecks. This supports smarter decisions and continuous improvement.

Vanie LLM’s role in transforming contact centers

Vanie LLM isn’t just a general-purpose AI—it’s built for contact centers. It avoids rigid rules, keeps learning from real conversations, and works across every channel. For organizations aiming to drive down costs, improve compliance, and build lasting customer relationships, it represents a major step forward.

Contact centers moving from outdated, rule-heavy systems to flexible, conversation-driven platforms like Vanie LLM set themselves up to meet today’s customer expectations—and gain a clear competitive edge.