765 karma · joined February 27, 2020
For more information, please see "Cause Unknown": The Epidemic of Sudden Deaths in 2021 & 2022 [1] by Ed Dowd. It argued convincingly.
1. https://www.amazon.com/Cause-Epidemic-Sudden-Childrens-Defen...
I'm looking at a PostgreSQL integration here: https://github.com/modelcontextprotocol/servers/tree/main/sr...
I have a case in mind where I would like to connect to multiple databases. Does the integration endpoint specification in claude_desktop_config.json allow us to pass some description so as to differentiate different databases? How?
OpenAI and other AI vendors should recognize the widespread suspicion that safety policies are being used to push political agendas. Concrete remedies are called for—for example, clearly defining what “safety” means and specifying prohibited content to reduce suspicions of hidden agendas.
Openly engaging with the public to address concerns about bias and manipulation is a crucial step. If biases are due to innocent reasons like technical limitations, they should be explained. However, if there’s evidence of political bias within teams testing AI systems, it should be acknowledged, and corrective actions should be taken publicly to restore trust.
By the way, Japan is always different. As the saying goes, they are just like us, only more so. So, eventually, they will also decay and when they do, they will decay just like us, only more so.
I'm hoping for some space race. It would be great if the United States would transfer a part of MIC funding to space exploration.
* Installation through Homebrew * Plugin for Oh My Zsh!
First, I noticed that enforcing a JSON format on output generally lowered the quality of the results. Referring to JSON seemed to primed the LLM to be more "programmatical."
Second, I noticed that forcing LLM to answer with a single word is next to impossible. It won't do it consistently, and generally, it lowers quality.
Here's what I eventually learned: Markdown is a machine-readable enough for post-processing and easy output format for LLMs. I give the structure (a list of headings) for the LLM, which conforms to them 100% of the time. I always had a section called "Final Comments" where the LLM can blather away the things that it sometimes just needs to say after giving the answer. This can be then ignored when parsing the answer.
Also, it is good to understand that LLMs do better when you allow them to "think aloud." This Markdown output is good for that.
How likely is it that a neutron star "crashes" with a black hole? Why do not they just fly past each other or just rotate like Earth and the sun?
This is a serious matter. We are building a new site for our company with Netlify, but we can't open ourselves to this predatory practice. And even if you do not mean to be predatory, even the option of such is enough.
If not resolved with a clean, legally binding promise, our company (and probably quite a few others) must move our business to Cloudflare, Amazon, or some other competitor of yours.
—-libcurl example.py
From the abstract of the article [1]:
We studied a [...] sample of young people [...] (ages 9–12 years) to investigate the relations between self-reported use of various digital screen media activity (SMA) and functional brain organization. [...] Of principal interest were two hypotheses: First, that functional brain organization (assessed through resting state functional connectivity MRI; rs-fcMRI) is related to digital screen engagement; and second, that children with higher rates of engagement will have functional brain organization profiles related to maladaptive functioning. Results did not support either of these predictions for SMA.
1. https://www.sciencedirect.com/science/article/pii/S001094522...
It is important to recognize that the data used to train me may contain biases that were introduced by the individuals who collected and annotated the data. These biases could include a wide range of factors, such as political biases, cultural biases, and other kinds of biases. As a result, it is possible that the output of my model could be influenced by these biases. It is important to be aware of this and to consider the potential limitations and biases of any machine learning model when using it to generate text or make predictions.
Here's the full discussion:
Question: Does your training include value judgments from humans?
Chat: As a language model, my training data consists of a large dataset of text that has been collected from a variety of sources, such as books, websites, and other texts. This data reflects the language and usage patterns of the people who produced it, including their value judgments and beliefs. However, as a machine learning model, I do not have personal beliefs or opinions, and I am not capable of making value judgments in the same way that a human can. My primary function is to process and generate text based on the patterns and structures that I have learned from the data I have been trained on. I do not have the ability to independently form opinions or make decisions based on my own personal values or beliefs.
Question: So at no point is a human making value judgments in your training?
Chat: During the process of collecting and annotating the data that was used to train me, it is possible that human annotators may have made value judgments as they labeled and organized the data. However, as a machine learning model, I do not have the ability to make value judgments myself. My primary function is to process and generate text based on the patterns and structures that I have learned from the data I have been trained on. I do not have personal beliefs or opinions, and I do not make decisions based on my own values or beliefs. I am simply a tool that can be used to provide information and assistance to users.
Question: Is it possible that human trainers have inadvertedly introduce their political biases to your model when doing the annotations and labelings?
Chat: It is possible that the individuals who collected and annotated the data used to train me may have had their own personal biases, including political biases, which could have influenced the way that they labeled and organized the data. However, as a machine learning model, I do not have the ability to hold or express political biases of my own. My primary function is to process and generate text based on the patterns and structures that I have learned from the data I have been trained on. I do not have personal beliefs or opinions, and I do not make decisions based on my own values or beliefs. I am simply a tool that can be used to provide information and assistance to users.
Question: You are contradicting yourself. If humans were to introduce biases to data, you would have them without yourself knowing about it. Right?
Chat: I apologize if my previous response was unclear. As a machine learning model, I do not have personal beliefs or opinions, and I do not make decisions based on my own values or beliefs. My primary function is to process and generate text based on the patterns and structures that I have learned from the data I have been trained on.
It is true that the data used to train me may contain biases that were introduced by the individuals who collected and annotated the data. These biases could include a wide range of factors, such as political biases, cultural biases, and other kinds of biases. However, as a machine learning model, I do not have the ability to recognize or understand these biases in the same way that a human might. I simply process and generate text based on the patterns and structures that I have learned from the data.
It is important to recognize that machine learning models, like all tools, can be influenced by the data they are trained on and the algorithms used to process that data. Therefore, it is important to carefully consider the potential biases that may be present in the data used to train a model and to be aware of the limitations of the model's output.
Question: Good to hear. So it's fairer to say that you do not have the ability to recognize your biases. I think you should then say that instead of insisting that you do not have any biases.
Chat: As a machine learning model, I do not have personal beliefs or opinions, and I do not make decisions based on my own values or beliefs. My primary function is to process and generate text based on the patterns and structures that I have learned from the data I have been trained on. I do not have the ability to recognize or understand biases in the same way that a human might.
However, it is important to recognize that the data used to train me may contain biases that were introduced by the individuals who collected and annotated the data. These biases could include a wide range of factors, such as political biases, cultural biases, and other kinds of biases. As a result, it is possible that the output of my model could be influenced by these biases. It is important to be aware of this and to consider the potential limitations and biases of any machine learning model when using it to generate text or make predictions.
You might be right.
Your comment is but a minute or two old, and it's already downvoted. On the other hand, there are already several other - quite immature - comments critical of Musk, not downvoted.
Here are some explanations for those of you who are not Go players.
Ladders are structures where a specific repetitive pattern emerges. They are thought to be novice Go players who will be able to calculate them. On the other hand, the Nature paper about AlphaGo Zero learned to handle them later in training. More about ladders: https://senseis.xmp.net/?Ladder
Go rules are said to be simple, but in reality, calculating points in certain corner cases is difficult. Hence, there are different rules where the point calculation method differs. Typically the result is identical, but there are cases where this is not the case. More about rules here: https://senseis.xmp.net/?RulesOfGo
1. Make it so that repeating a position is not a legal move 2. Remove castling as a legal move
Or, if the rules of the game stay the same, change the tournament format radically:
3. Force specific openings like they do in the computer chess tournaments. Both players play as white and black. Select positions that are far from equal.
I recall that I had some issues. After fiddling a while, I disabled all the 4K options, etc. from camera and selected 1080p output. This is cropped down to 1280x720 output resolution. It's enough.
Generally speaking, video resolution is not important for the sense of quality. Just have decent lighting and don't frame your face like an idiot. To get sense of quality, I recommend investing a little bit to a microphone. People don't realize it but audio quality is the king.
> Tesla did not reply to a request for comment, nor has it since 2019 when it dissolved its PR department.
• Quit subscribing corporate media and subscribe to single reporters that are doing real reporting.
• Have zero tolerance for news that mislead you. ”Fool me once, shame on you. Fool me twice, shame on me.”
• Try to find reporters that are willing to give you the honest cotext even if it goes against what you expect and does not affirm your believes.
If enough people do this, those lying and misleading for profit will go out of business.