About Our Broken Politics, Broken Media, Broken Republic

Search engines

Until recently this was by far the best way to do good research. Google, Bing, DuckDuckGo and others present similar, if not identical search results. All of these engines claim to target advertising but not search results. Search terms hold the key.
If you search for "trump 17 trillion investment" the results will be similar (and disturbing). The results will be a mix of real facts and stupid opinions. Be careful and verify.

Fact checkers

"You can have your own opinion but not your own facts,"
There are many fact checkers around the world. The Poynter Foundation is an institution that registers fact checkers that meet a code of ethics.
Signatories of IFCN Code of Principles
IFCN Code of Principles

These meet the IFCN Code of Principles and are signatories:
- snopes.com
- factcheck.org
- politiFact.com
- reuters.com/fact-check/
- apnews.com
- checkyourfact.com
- leadstories.com
- allsides.com is not a IFCN signatory but is a useful reference for a conservative view.
- npr.org is not a IFCN signatory but is a useful reference for a liberal view.

This is not to say that these sites are completely unbiased. It is useful to compare sources and look for consensus.
It is notable that Fox News, OANN, Newsmax, The Daily wire, The American Conservative, The National Review and Breitbart do not have fact checkers.
It should also be noted that the parties and people in power will receive the most attention from the fact checkers. This does not imply bias, but is simply a focus on the most important issues of the day.

What is our confidence level?

It is a very human thing to be overly confident. With humility we can apply a confidence percent to the stuff that we believe. 100% confidence is almost impossible.
A video about confidence

Who is it?

If it's a non-profit institution or a think tank, it is useful to look up who provides the funding. Nonprofits must have written policies in place to manage and prevent conflicts of interest. Nonprofits are, in general good sources of information but can be biased, be aware.

Government sources

Government sources at the local, state and federal levels are all generally good sources of information, but can be biased by the executive branch and other influences. The most reliable government information is that which is needed for a healthy economy such as weather reports, economic performance history and foreign travel advisories.
Some care must be taken however with short term economic data due to the voluntary data collection methodology.

Online and broadcast media

It must be understood that every media source selects its target audience and as a consequence it will not offend that audience. This inevitably leads to bias. This is not to say that good information cannot be found from these sources but only that we must be aware of the bias. The best way to counter bias is to seek consensus among many sources.

Social media

Unlike online and broadcast media, YOU are the only audience on social media. Every bit of content is designed to keep you engaged. The terrible result of this is Confirmation Bias.
Social media is great for keeping in touch with friends and family but is horrid as a source of information. It is the single most prolific source of lies and misinformation in history.
If you need an ego boost from LordBebe21 on X, go for it, but if you want to find good information avoid social media at all cost.

Video, pictures, data and postings without provenance

YouTube and other video sources comprise a large trove of useful information if one is careful. Too many times when a video or picture is posted as a current event it is actually from a different place and time. The source of videos and pictures are often not indicated.
An image search is useful.

The Good and bad of A.I.

A.I. is without a doubt the most consequential invention ever in human history. But it is not what people think. A.I. is better thought of as a computer program of a fantastically intelligent newborn baby. The intelligence and memory of this baby is only limited by hardware resources. This baby must be taught, what developers call "training" and this is where the rubber hits the road.
Whereas a human baby will be taught by parents, siblings and exposure to the real world, these large A.I. models are being trained by unmoderated data from a variety of sources. The parents of a human baby will teach with love and hope, and some sense of ethics from real-world experience. The poor A.I baby has none of that, and is impaired as a result. For example a human baby is not born racist but can certainly be schooled to be so, and A.I. can as well.
A human parent will attempt to correct a child's bad behavior through coaxing or punishment. The A.I. developer will similarly attempt to correct A.I. bad behavior using top level system prompts that are invisible to the user.
To date the owners of these large A.I. models are concealing what data is being used to train the A.I. and what the top level system prompts are as "proprietary" information. In addition there is no requirement by law to clearly indicate that content is A.I. generated. This is dangerous.

That said, small A.I. models trained with specific data could be an astoundingly useful tool. Imagine an A.I. trained on the successes and failures of designing and constructing a bridge. It could be trained using the breadth of state-of-the-art engineering and science. The result would be a cheaper, faster and more reliable structure. This is where A.I. excels.

The commercially available A.I. today such as ChatGPT, Claude, and DeepSeek are useful with caution. An A.I. query is nothing like a Google or Bing search however.
Google:
• You use keywords
• It returns links
• You evaluate sources yourself
• Vague queries often work fine
A.I.:
• Will respond to full questions and context
• Will generate answers directly
• Should cite sources (and you should verify them)
• More context = better results

What Helps A.I. Give Better Responses
1. Be Specific About What You Want
• ❌ "Tell me about OSHA"
• ✅ "Give me the benefits and criticisms of OSHA using only government sources, in XHTML format"
2. Tell AI Your Constraints
• "Use only official government sources"
• "No sources from far-left or far-right organizations"
• "I need this for a center-right audience"
• These dramatically improve relevance
3. Specify Output Format
• Such as: "Output in XHTML format named Fubar.xhtml"
• Also useful: "Give me a table," "Make a bulleted list," "Write this at an 8th-grade reading level"
4. Tell A.I. Your Purpose/Audience
• "I'm creating this for my conservative friends who are skeptical"
• "I need to fact-check an article"
• This helps me calibrate tone, depth, and framing
5. Correct A.I. When It Misses the Mark
• "That's too partisan—try again with neutral language"
• "I need more sources"
• "Too technical—simplify"

Where A.I. falls short
If information is sparse on a particular subject A.I. will still attempt a response. This results in the well noted ability of A.I. to "hallucinate". The user must be aware of this fact.
Groc was recently contaminated with top level system prompts that provided extreme right wing responses when non-specific queries were made. This has since been corrected as claimed by X, but we should be aware of this when using ALL A.I., BE SPECIFIC!

Proportion and equivalence

"A billion here, a billion there, and pretty soon you're talking about real money" from Senator Everett Dirksen in the sixties, is cogent today in some contexts. A billion dollars is of course a lot of money to an individual but is pocket change to an economy the size of ours. It works out to about $6 for every American taxpayer, not much to be concerned about. A billion dollar project here in Clark county however would be about $7100 per resident, and that would be momentous.

Every day numbers are thrown around without proportion, equivalence or context. Without doing the math we have no idea how important a number really is.
An example of this is our national debt and the interest on that debt. It works out to be about $112,000 for every man, woman and child in the US. The interest on this debt is equivalent to about $5600 for every man, woman and child in the US every year. This interest goes directly into the pockets of the investor class. Are you panicked yet?

Correlation vs Causation

This is problematic across all of the media, particularly with respect to science. The public dispute about the cause of autism is a good example.
Autism correlates well with the increase of computer use. It also correlates well with the popularity of rap music. So do these things indicate the causes for autism? Maybe, if a mechanism could be found that links them to autism.

The above is indicative of the difficulty in understanding correlation. The problem here, is that many causes could be contributing to the rise of autism and until one or more biological mechanisms are discovered, we cannot tease out which one/s are to blame.

Be careful with definitions

It is helpful to look up the definitions of commonly used words and phrases.
Government vs business - Our government is a non-profit tasked with serving the American people, all of the People, per our Constitution. A business only survives on profit. Our government is not, and should not, be regarded as a business.

The terms "liberal education", "liberal democracy" and "liberal politics" mean completely different things.

The terms "socialist economy", "socialist democracy", and "socialist communism" mean completely different things.

In our Constitution the words "People" and "citizens" are distinctly different. The "People" refers to all people within our borders and the basic rights that they are granted. The word "citizens" refers to illegibility for participation in our government, including voting.

Some general rules

Finding out that you are correct gets you nowhere
But finding out when you are wrong gets you closer to reality. This should be an axiom for all of us in the post-truth era.
- Look it up
All manner of crap gets in the media. verify!
- Anecdotal accounts
In a country of 340 million almost anything will happen. It it significant? Is it a one-off event or is it a big deal?
Be suspicious of one-word quotes
Many times an article will be pages of opinion but few actual quotes. Context is everything. Search for the transcript if it's a speech, press conference or other verifiable text to find out what was actually said.
Be aware of overgeneralization.
If you see that 20% of the people at the beach are blonds, don't assume that 20% of the general population is blond.
A recent example is a headline like "Illegal immigrant kills landlord". The knee-jerk reaction is to have suspicion of illegal immigrants. You will be mislead without looking up actual crime figures.
Saying something over and over doesn't make it true
Trump did not win in 2020 despite how many times he says it. He can say this because we have a free country, but courts require evidence that he does not have. He lost them all.