You know you can just not use AI, right?
You can make a choice not to be part of this.
Even if your job uses it, YOU don’t have to in your normal life. You don’t have to let your kids use it.
You didn’t have it three years ago. You can just…be on Team Human. You can choose.
@Catvalente That's not true.
A major danger of algorithmic abuse has been projected for years to be the risk that an uncaring government reform might replace trained bureaucrats with simplistic algorithms, and use their suggestions to mete out government power on non-consenting population whether it consents or not. Despite the warnings, this sort has even happened already, too — for example, in systems designed to mask racism by having a computer express the stereotype that Black people are offensive and criminal.
I agree we should critically evaluate our personal use of AI vapourware, understand practical uses of Machine Learning (ML), and also understand existing abuses of statistical algorithms for decision making.
In that quoted article, I didnt see any reference to Bayesian reasoning or the models which are likely to be involved too. The people writing these articles for the general public are often just reiterating a historical statement like "it's too complicated to explain". The problem is that many people cannot grasp the use or misuse of statistics in law or policing, whether it is informed by ML or not. The difference between a statement about where general crime or crime types has occurred historically and the evidence needed to prove that a specific person is committing a new crime often escapes police, witnesses and judges. They are totally separate concerns. Calling any of this AI is just another layer of confusion.
yes, its easy. moreso if you already questioned the cloud.
@Catvalente or you selectively use it for the cases it was designed for when you want to deflect liability onto a machine...
I really don't know why anyone would accept a "sorry the AI must have gotten it wrong and I didn't notice" as an get out of jail free card but enough apparently do...
@Catvalente More like Team Not-LLM!
The opposite of "AI" is not "human".
Or just use you AI locally 🦾 💻 🧠
I completely understand the concerns about relying too heavily on AI, especially cloud-based, centralized models like ChatGPT. The issues of privacy, energy consumption, and the potential for misuse are very real and valid. However, I believe there's a middle ground that allows us to benefit from the advantages of AI without compromising our values or autonomy.
Instead of rejecting AI outright, we can opt for open-source models that run on local hardware. I've been using local language models (LLMs) on my own hardware. This approach offers several benefits:
- Privacy - By running models locally, we can ensure that our data stays within our control and isn't sent to third-party servers.
- Transparency - Open-source models allow us to understand how the AI works, making it easier to identify and correct biases or errors.
- Customization - Local models can be tailored to our specific needs, whether it's for accessibility, learning, or creative projects.
- Energy Efficiency - Local processing can be more energy-efficient than relying on large, centralized data centers.
- Empowerment - Using AI as a tool to augment our own abilities, rather than replacing them, can help us learn and grow. It's about leveraging technology to enhance our human potential, not diminish it.
For example, I use local LLMs for tasks like proofreading, transcribing audio, and even generating image descriptions. Instead of ChatGPT and Grok, I utilize Jan.ai with Mistral, Llama, OpenCoder, Qwen3, R1, WhisperAI, and Piper. These tools help me be more productive and creative, but they don't replace my own thinking or decision-making.
It's also crucial to advocate for policies and practices that ensure AI is used ethically and responsibly. This includes pushing back against government overreach and corporate misuse, as well as supporting initiatives that promote open-source and accessible technologies.
In conclusion, while it's important to be critical of AI and its potential downsides, I believe that a balanced, thoughtful approach can allow us to harness its benefits without sacrificing our values. Let's choose to be informed, engaged, and proactive in shaping the future of AI.
CC: @Catvalente @audubonballroon
@calsnoboarder @craigduncan
#ArtificialIntelligence #OpenSource #LocalModels #PrivacyLLM #Customization #LocalAI #Empowerment #DigitalLiteracy #CriticalThinking #EthicalAI #ResponsibleAI #Accessibility #Inclusion #Education
Let's not think that running locally solves the centralisation poblem - unless you roll your own model you can't know what is in it, what it was trained on and it remains a privacy risk for anything sucked into it. Even if it works for some uses, it is very hard for any of us to understand the risks here and I definitely wouldn't trust any of the businesses creating models - to do the right thing - at this time.
It's true that simply running models locally doesn't completely address the issues, and it's important to be aware of the potential risks. How do you balance the benefits of LLMs with privacy concerns? Do you avoid using them altogether?
I believe that utilizing open-source models developed by communities, such as OpenCoder and Piper, is a step in the right direction. These models can be run locally and have open, reproducible training processes, giving users more control over their data and reducing privacy risks. I trust the FLOSS community, and using open-source models aligns with my values.
Additionally, I run and use LLMs offline to mitigate most privacy risks. By not connecting the LLM to the internet, except when necessary for specific tasks, we can minimize exposure to potential security threats.
You're right that unless you create your own model, you can't be sure there are no potential privacy risks. However, don't you think the risks posed by LocalAI are acceptable? Of course, self-training an LLM is the optimal way to achieve the highest level of control and security. While it may not be feasible for everyone due to current resource constraints, I'm hopeful it will become more accessible in the future. I'm considering it myself, but with the current economy, investing in an LLM server seems unreasonable for me. I'm curious, do you have experience with it?
@debby
Privacy is one issue but there are many downsides to using LLMs. I won't open all that up, but running locally is a minor mitigation is the main point I'm making. It achieves little while reducing the effectiveness of LLMs due to reduced processing and memory.
I use them locally occasionally but have not found them more useful than a web search except for a handful of small tasks, so for me quite a lot has to improve.
@debby Unfortunately, it's a misnomer to refer to most of the LLMs that you can run locally as open-source. The code that runs the inference is open-source, but the model weights are a big inscrutable blob, and the training data and process for those are usually proprietary. So the usual concerns about the training data (unauthorized use of copyrighted work) and process (exploitation of underpaid labor) still apply.
tell me you used an LLM to write this without telling me you used an LLM to write this.
@calsnoboarder
LLMs have completely solved for the "minimum word count" problem for HS students & undergrads.
What they _haven't_ done is demonstrate any useful thinking. Solving for form without solving for content: the essence of bullshitting.
Unfortunately, it interacts badly with the management types, who often _use_ form to evaluate content -- "aha this is well-punctuated, has bullet points, & I understand part of it; must be right"
@Catvalente
I find it amusing to think in terms of replacement—like when a message I type is read aloud by an AI voice. Do you think I'm being replaced? I don't see it that way. Instead, I think of it as making things more accessible. When you type a message and a screen reader with an AI voice reads it aloud, I don't see it as a replacement; I see it as a tool that makes things more accessible.
In fact, I often use voice typing instead of manual writing because it's faster and less tiring for me. With the help of an LLM, my transcript is spell-checked and proofread, and then read back to me in my own voice. I love this process—it's a simple and useful tool that reduces fatigue and makes writing more comfortable.
For me, AI is a liberating force that helps me express myself more freely and with less effort. It's a tool that enhances my abilities rather than replacing them. And I think that's something to be celebrated.
So, am I being replaced?
Every use of AI sends money & manipulation to its investors; the fossil fuel industry, which is desperately fending off an urgently needed phase-out by funding wasteful uses of energy.
https://www.bloomberg.com/news/articles/2025-04-15/ai-to-prop-up-fossil-fuels-and-slow-emissions-decline-bnef-says
https://www.theatlantic.com/technology/archive/2025/02/ai-energy-crisis-fossil-fuels/681653/
https://www.technologyreview.com/2025/05/20/1116272/ai-natural-gas-data-centers-energy-power-plants/
https://www.popsci.com/technology/ai-fossil-fuels/
Stop helping petrostate despots like MBS, Putin, and Koch Network, they're frying the planet.
MFDoomer
Catherynne M. Valente
Riley S. Faelan
Craig Duncan
relogi
calsnoboarder
Claire Simpson
Amber Grey
lin11c
cognitively accessible math
CJ Paloma In Monsoon Season
Becca Cotton-Weinhold
Flesh
Klaus Frank
AstraLuma
Pesce MediaWorks
Big Ducky
mirabilos
cholling
Frost, glow wolf 🐺
Mad Alex
CyclesSmiles
smallerdemon
Mad A. Argon 
Paul SomeoneElse
Hudson
Barry Cook
�
Debby
happyborg
Matt Campbell
Jeremy Kahn
Asta [AMP]
Mikel Forcada EA5IYL
Nicole Parsons
Pete Alex Harris🦡🕸️🌲/∞🪐∫
Ω 🌍 Gus Posey
Unknown Pseudoartist