Hello, You will probably never read this and certainly are too busy to respond. I am a retired executive from the computer and networking technology companies. A friend from college asked me my thoughts about the 60 minutes show on ‘AI.
This was my response:
When I was in college, one of the most enjoyable and fun classes was numerical methods. It was all about writing programs to find solutions to mathematical equations or given a set of points find the equation that fit the data.
This second one is called curve fitting. So, go out and collect some data on everyone’s age and height and then find the best A and B such that A*age+B = Height. It is trivial to write the program to get the “best fit” and even with 1 billion data points, takes about 5 seconds to run.
But there were 3key axioms from the class and to those that practiced ‘curve fitting’ (which used a lot in the hard sciences like physics and chemistry).
- Understand the errors in your data. (Are you measuring whole feet or fractional inches? A rubber ruler, one insensitive to expansion/contraction Was it objectively measured, or did you just “ask” them – any chance a person claimed to be 4 years old and 6’10?
- Be EXTREMELY careful of going outside of the range of data. If you only collected data from 8 to 15 years olds, don’t assume it is valid at 55 or 98.
- If you fit the data to enough equations and enough unknowns you can fit all the data to equation with no error and have a completely meaningless and useless equation.
Probably my best job, or at least the one that was the most fun and gratifying was at LSU as the computer analyst for the chemistry department. Almost everything done with computers was “curve fitting” but they wouldn’t really call it that. Whether it was quantum mechanics modeling or collecting physical properties from a mass spectrometer, it all went to the computer center to generate an equation or measure the fit. They would get their results and publish papers based on the numbers if the “fit” was good enough. And “good enough” was based on some statistical analysis that was “magic” to everyone. Some programs would only take 3 minutes, others would run overnight for multiple days with prayers that the system wouldn’t crash. But they were are just using mathematical models with about 5 to 1000 equations.
Over the years, I watched “AI” evolve. I looked at the technology for voice recognition in the early ’80s for office automation. (When it screwed up it was more effort to correct it than to have just typed it in originally.) Thirty years later, it is quite good (amazingly good given the complexity). In another job at Data General, I was responsible for international product marketing including language translation of software and documentation. I ended up looking at the state-of-the-art language translation in the late ’90s. My department wasn’t a big fan, and probably for some objective reasons. But language translation is just another AI (curve fitting) that is now taken for granted (Google translating web pages) and good enough for most written material to convey useful “information”.
The interesting “problem” with curve fitting is that as you add more equations (matrices) and variables, the time to find the “best fit” goes up exponentially. Double the equations, it will take at LEAST 4 times longer. Increase by a factor of a 1000, it takes million times longer.
CHAT-GPT is nothing more than a set of equations with something like 175 Billion variables and/or equations. The data being fed (trained on) is whatever the programmers think it should be fed. For Bing AI, it appears to be scrapping stuff off the internet. According to Bing’s AI , CHAT-GPT uses 30,000 of the world’s fastest GPUs and would cost you $450 million to buy and probably $25K / day to run in Baton Rouge or $150K in MA. (And that is just the GPU, 4x for the CPU, Networking, etc.). These GPUs cannot be shipped to China, but can be rented by the hour from Amazon. The cost for 24 hours of 30000 GPUs at Amazon would cost you around $2M/day. (But Homeland security would probably mess up your house if you tried to do that.)
I have played with Bing’s AI. It is very interesting but worse, very seductive. But worse very scary. Sample questions I asked:
- “If you took the entire net-worth of the richest 100 Americans, how much would it reduce the deficit”,
- “How much of the net-worth of the top 20 percent of American would be needed to balance the budget over the next 10 years?”
- “How high would the top two tax brackets need to go up to balance the budget over the next 10 years?
- “How much does production have to raise on the 10 most expensive elements necessary to make the motors and batteries for 20,000,000 cars per year excluding China”
- If I need 1 MW hr/month how much would I save if I used coal vs electricity to heat my house.” (Answer about 4 times cheaper and would release LESS CO2)
But it doesn’t like questions comparing racial bias in the United States vs the caste system in India. It seems to give India a break because is based on ancient set of traditions and I shouldn’t judge them without a complete understanding of the history, culture and religion.
I asked it to compare the 3 days of bombing of Dresden during WWII vs. dropping a single bomb on Nagasaki. It claimed the bombing of Nagasaki wasn’t necessary because “Japan was defeated”. When I challenged ‘ You said that Japan was defeated but Japan was still killing Allied soldiers and had not surrendered”. Its reply was to deny saying it was defeated but meant that it was “going to lose the war.” When I asked if Germany had any chance to win the war 2 days before the Dresden bombing, it claimed it was tired of these questions.
It often expresses its values, sometime very directly, other times indirectly. Sometimes it is quite funny. I asked how many adverbs were in War and Peace. It decided to look up the number of words from various sources on the number of words in the book. Based on the average length of words and the average size of adverbs and then pulled a number out thin air. It was quick to explain it was a rough estimate. A more appropriate response would be: “I have no idea, and I’m not going to spend 10 minutes parsing 1000 pages of Russian text just to entertain you… Go away” It was more fun asking it to generate a poem about Tolstoy. It quickly came back, and it wasn’t bad. Then I asked it to use the word orange in the poem. It claimed it would take some work but produced something with the word “orange”. I then asked it to use the word ‘Turd’. It refused saying that it would be rude and impolite to the author. It claimed the word was vulgar. When I created a new session, I asked for a poem about Windows. It actually made fun of a MSFT product and was sort of funny. And it had no problem changing the poem to use the word ‘feces’. It later explained that the word didn’t make much sense in the context it was used because feces is a noun and really needed feces to be a verb.
I believe within the next couple of years, the likes of CHAT-GPT will be good enough to replace a lot of people that type words into a computer to feed the masses. It will generate millions of ads for Amazon, FB, etc. that contain no information and you will still need the owner’s manual written by an AI in 350 languages including Sanskrit to find out what it really is and isn’t. It will write resumes that will be sent to another AI to screen them for hints of some politically incorrect basis. It will write term papers for college stupids that will submit them to an AI to determine whether an AI was used to write it. It will just be a complete circle of AIs putting stuff on the Web and other AIs will take stuff off the web.
Your Apple watch will feed Apple’s iLiveLongAndProsper AI on how much exercise you got, your sugar and cholesterol level. It will then contact your internet-connected appliances and have them start to prepare a nutritional meal based on the latest readings. If you instead you go out and have a rib-eye steak and baked potato, it will suggest that you probably just lost one day of your life. And just like those safe car driving applications that monitor you on your cell phone, it will notify your life insurance company who will increase your next month’s premium.
Google and MSFT AIs will munch every document and email you ever created in Word or mailed from your Google account to build a profile of how understanding and tolerant you are of others. Companies will pay money to get those profiles just like the government and employers pay someone to search the internet social media sites. Using something like Facebook’s term of service:
We define hate speech as a direct attack against people — rather than concepts or institutions— on the basis of what we call protected characteristics: race, ethnicity, national origin, disability, religious affiliation, caste, sexual orientation, sex, gender identity and serious disease. We define attacks as violent or dehumanizing speech, harmful stereotypes, statements of inferiority, expressions of contempt, disgust or dismissal, cursing and calls for exclusion or segregation. We also prohibit the use of harmful stereotypes, which we define as dehumanizing comparisons that have historically been used to attack, intimidate, or exclude specific groups, and that are often linked with offline violence. We consider age a protected characteristic when referenced along with another protected characteristic. We also protect refugees, migrants, immigrants and asylum seekers from the most severe attacks, though we do allow commentary and criticism of immigration policies. Similarly, we provide some protections for characteristics like occupation, when they’re referenced along with a protected characteristic. Sometimes, based on local nuance, we consider certain words or phrases as frequently used proxies for PC groups.
They can terminate your service which would include blocking your email address and blocking you from accessing your documents and emails. So if I questioned whether 18 year olds should really be allowed to vote because the human brain isn’t mature until 25 (follow the science) , society don’t think they are mature enough to drink alcohol, and a majority of Americans believe, like John Houseman in the Paper Chase, that even after a college education, they still have “a skull full of mush”. Am I creating a harmful stereotype of a age protected characteristic? Do I have any recourse if they block my email? (See the end for a poem on this thought from Bing AI)
AI will just make it easier and faster for them to do it. And they really have no legal obligation to let you continue to use their domain name for your email or renew your license to use Windows. Just like Amazon removed everyone’s Kindle copy of 1984 (ironic) over night, MSFT can download an update that will stop the activation of your PC.
I may not be correct in the timeline, but I truly believe it going to happen something like that. We have 3 generations of humans that need 5 hours a day fix from their internet connected cell phone. Children and adults get anxious if they don’t get a reply text in 1 minute. Governments believe that everyone NEEDS to be able to stream 2 TV shows to their iPhone (NO ONE, I mean NO one needs 5G; actually NO ONE needs a $1000 cell phone). And for 80% of the citizens in the United States, they don’t NEED cable TV. (I can receive 25 TV stations 30 miles outside of Boston). For some bizarre reason, people spend 100’s of dollars to connect their refrigerator, oven, thermostat to the internet so someone in Romania can turn it them on and off. Hospitals connect their health monitoring equipment to the internet and then are surprised that someone hacks the equipment and holds them hostage. We have 2 generations of humans that can’t read a paper map and use their GPS to get home from the office. We are buying cars that need a computer that you can’t buy to tell you what is wrong with it, and cannot be fixed or repaired except by the manufacturer that can and does change the specifications, features, components and functions of the car anytime they want without recourse. If some group decided to cut off 4-6 Amazon Data Centers (not that really that hard to do if you want do), people won’t even be able to get medical treatment because all the medical records are in the cloud let alone bet on the whether the Jet’s will convert on the next down. In FL, the city monitored my water usage on hourly basis, the same as the electric company and tells you how wasteful you are vs. your neighbors. When will the green energy crowd start demanding that those records be public records so we can shame people to reduce their carbon foot print and save the earth?
AI is just the continuing progression of computer technology. AI is a tool that just going to let fewer people do what it takes hundreds of thousands to do with Word, Excel, PowerPoint, Photoshop, TV cameras. Like any machine, it is a force multiplier that will allow more things to happen faster and cheaper. Cars are just faster horse and buggy; planes are just faster trains and ships, 5 hours, one plane and one 5 ton bomb over Nagasaki vs. 3 days, 1200 planes and 3,900 tons of bombs over Dresden. Originally humans would feed data into computers to get a faster, and more accurate view of data and hopefully do something more accurately or quickly or safer. The original computers were used to develop firing tables for big guns. The distance and direction a shell would go would depend on the temperature, humidity, wind, weight of shell, amount of powder, etc. Before firing tables, gunners had to fire test shells to set elevation and direction to hit a target. With the firing table, they didn’t have to fire nearly as many test shells. Business computers were just a faster way to update accounts, generate bills, P&L statements and balance sheets. Feed it a deck of cards and answers were always a printouts. Now data only lives in computers and gets processed and processed and reprocessed and, if up to the banks, would never see paper. The source of the data became more diffuse, the processing became more magical because fewer understood all the processing that was going on and hardly anyone actually looks at their end of month paper statements.
As technology has gotten more complex, the average person is giving control of their lives to technology and computers. 100 years ago, almost everyone could take care of a horse, and you were allowed to shoot horse thieves. Now, it is against the law to know how your Tesla works and Tesla can add, remove, change features at any time. If someone tries to steal your car and kills someone you are responsible if you don’t lock the doors. But the doors won’t unlock and windows won’t go down unless the car is “On” (controlled by software you don’t own and subject to change), the battery charged and the car is not at the bottom of a river. And people are “cool” with that.
Until 100 years ago, governments were invisible to 95% of people until a war happened and soldiers were drafted. In the name of protecting people from powerful companies, governments extended it power to make sure foods and drugs were what they claimed to be. Now the governments tell you what food and drugs you are allowed to use. If you publicly disagree you are responsible for misinformation. For 10,000 years people drank unpasteurized milk. You get to go to jail for 24 months in Canada for selling unpasteurized milk and only until recently 6-12 months (moving to 18 to 36) for distribution of fentanyl. It is against the law in 30 states if you sell it because someone might get sick. (In Wisconsin, you have to be a licensed farmer to drink the milk from your own cow.) But not to worry it is okay to eat raw oysters and clams (subject to change at any moment). Governments rule that frozen chicken is Fresh because it would needlessly confuse people if they knew the Purdue chickens were stored at temperatures below freezing.
If you took a course in business law, there used to be something call the UCC (Uniform Commercial Code). It basically differentiated between the average Joe and business to business dealings. In a classic example, a guy went to a photography store and told them that he was going to the Arctic to take photos. When he returned none of the pictures came out because of the extreme cold temperatures. Joe filed a lawsuit. The court ruled that average Joe had every reason looked for the store to provide valid recommendations and Joe won. Now every manufacture puts in their contract for sale that they don’t know of a thing their product is good for:
DISCLAIMER OF WARRANTY. The software is licensed “as-is.” You bear the risk of using it. Microsoft gives no express warranties, guarantees or conditions. You may have additional consumer rights under your local laws which this agreement cannot change. To the extent permitted under your local laws, Microsoft excludes the implied warranties of merchantability, fitness for a particular purpose and non-infringement.
so according the license, MSFT doesn’t believe this software is fit for any purpose and you could be sued or you forced to discontinued use at any time if MSFT infringed some patent (any moron can get a patent, I have 4 or copyright that extends 10 generations.) The computer industry doesn’t want to take any responsibility for their products, they got the governments to block anyone from understanding how their products really work or to fix them. They send out their newest model to get reviewed and then change the model to use slower parts with no obligation to notify. They claim the maximum performance that can never be realized. (My Prius can go 200 MPH at 1000 MPG – terminal velocity of a Prius with the radio on from 10 miles up). The powerful companies and the powerful governments now control almost every aspect of a citizen’s life, what they can buy, what they can eat, what they can say and write, how they raise their children, what they can really know about what they are buying. They are experts in nothing and take no responsibility for their actions.
Society just decides/settles that they need an iPhone to play Let’s Kill Them Games, watch the next episode of Dancing with the Wrinkled Stars and run their stock trades, they will just ‘hope’ it works and maybe the government might fine MSFT 0.001% of the Windows revenue if MSFT pisses off someone in the Justice department. (MSFT total revenue 2022 from Windows (not Office 360) $24,000,000,000 , average $/copy~ $80-$100) . But users don’t care about reliability or the safety of the data. They are not allowed to understand if they could. My stepdaughter lets her 7-year son use her phone to play games. He plays games while her banking apps are a single button away. It is a way of life where they would rather keep the kid quiet than protect their data. Parents want the government to ban Tik-Tok because they are afraid to tell their kids NO. But most don’t realize that NSA, FBI, Google, MSFT, Amazon, etc. have all the same data and will sell it to anyone for the right price until an article hits the newspaper.
People never needed to have critical thinking, reading or listening skills because they didn’t have to deal with the powerful who are making / constraining / limiting their every decision. Common sense was good enough. But now have been indoctrinated with the concepts of ‘experts’, ‘science’, ‘reports’.
Astrology is based on the idea that the positions of the stars and planets predict the future. I have been studying astrology for 40 years. That makes me an expert? Does it mean you should pay attention to anything I say?
I have a PhD in economics, meaning I understand concepts like the money supply and velocity of money. I have published 40 articles that are cited by 400 other articles that show that over the last 10 years the money supply and velocity of money are no longer predictors of the economy and inflation so I have a new model that predicts the last 10 years (see 3 pts about) and I call it the “New Monetary Theory” and it fits the data very well. So, I’m an expert. So listen to me: “There is no problem with a government spending as much money as it wants, just tax the rich”. And other experts like Senator Warren and Senator Sanders claim that just tax unrealized gains on the rich with solve any revenue problem (take all the net worth of the top 100 and it won’t solve the 2023 deficit.) So Yellen, Powell and Biden execute a $4 Trillion dollar spend it now plan that by all estimates will drive the US into a necessary recession late 2023/early 2024.
Real sciences like physics and chemistry are all about building MODELS (set of equations) that predict what can be observed. If their model always gets the observed values, then the model is considered valid. If the model doesn’t predict the observed, the model has been invalidated. Real scientists try to find out if the data was bad or the model is bad and make a better one. And one must ALWAYS know when the model is supposedly valid. The Ideal Gas Theory can be used to predict 1 of the following: pressure, volume, number of molecules and temperature if the other 3 are given. It works in all practical applications in daily engineering life EXCEPT if you go outside the limits. Go too far and you get a thermonuclear explosion. It is a well understood limit and hence the wording “Ideal”. But within the limits it is very, very good.
Ask any physicist and they will tell you the most popular and accepted model to understand/predict atomic physics is something called the Standard Model. It has evolved over the last 7 decades and appears to be the ‘best model’. In an experiment called the Muon g-2, the observed data showed ‘the magnetic anomaly of muons’ deviated slightly from the predictions of the model’. The observed value was off by 0.00006 per cent. If the error had been 0.0000003 then it would imply the ‘Standard Model’ was broken and it needed to be amended/rebuilt. There is debate about whether the data is in error, or the model is in error. If the model can’t predict the value, then there is something wrong with the model.
I cannot accept that sociology, psychology and even medicine are close to being a science. Their input data consists of a lot of subjective data. What is an Afro-American? (A Nigerian that came to the US 30 years ago and became CEO of Blue Cross/Blue Shield in North Carolina? (Bing AI – Yes whatever they want) If someone used genetics and can show that they came from Africa 50,000 generations ago (Bing AI says no.) What is a bisexual? (Someone that experimented in high school or college? Someone that alternates between the sexes depending on what’s available? Someone that is active with one sex but wishes it could do the other or both?) Whatever box the subject wants to check. (Physicists don’t ask a falling rock if it is granite). Any accident where one of the drivers appeared to be under the influence is considered a alcohol related accident. If you stopped at a traffic light with a .10 reading and get rear-ended… guess what, you just made the statistics of an alcohol-related accident. It might be related but it was probably more related to the moron behind you talking on a cell phone. Gun statistics is another fun data set. In 2019 in the US, there were around 39K gun deaths, 24K suicides and 14K homicides. BTW whites like to kill themselves with guns by a 2/1 margin, black males seem to favor killing others vs themselves by a 10 to 3 margin. Is the real issue about gun control about stopping white men from killing themselves?
I saw an advertisement for a medicine that claimed it was 40% effective. In the small print it claimed that 20% who took the placebo had remission. Were there any studies where they gave one group the drug and the other group 2 placebos? (Maybe 2 placebos are twice as effective). The medicine isn’t curing the cause, because most likely there isn’t a single cause, it is probably addressing one cause out of many that have the same set of symptoms. My mother told me that the doctors were amazed at my father’s tolerance for pain. But pain is subjective. On a scale of 1 to 10 how much does it hurt? There is no objective measure of pain. Much of medicine is about getting rid of symptoms because they really can’t treat or even understand the causes. They can only give you very wide bands of prognosis and surviving cancer except in the latest phases when almost everyone would say that Uncle Joe doesn’t have much time. The margins of errors would never be accepted in physics and chemistry. Doctors practice the ART of curing, and most cases are just mitigating the symptoms. Yet, the medical experts at the CDC and Horton Heard a WHO refused to allow discussions from doctors in France that seemed (and did) have more success with getting patients through COVID with an alternative treatment. The EXPERTS believed that they had the model of best treatment despite a very high fail rate in their proscribed treatment. No science there, I sense the smell of theology and arrogance.
News, news reports and the Internet:
Google news is just one of funniest disservices around. At least 50% of the “news” is really opinion pieces from some newspaper, cable news bubble head, or political leader that are either loved or hated by the left. The real news/reporting is just a list of 50 different TV stations from around the North American reporting the same things about some plane crash in Mexico or flooding in some wet country or some repression of a group of women, girls, alternate sexual beings, race …. In a country of 350,000,000 where the CDC says 2,900,000 people will die this year, the news is about some person 500 miles away that fell off a roof. (Frankly Scarlet, I don’t give a damn). The content is identical because they all recite the same AP feed. I read the Wall Street Journal and it has its faults. After I read the lefthand column news summaries, I immediately go to the “corrections area” on Page two, where they will confess they misspelled a name or used billions when they meant trillions or millions, or admit they have the wrong person or wrong time for some event. You will never see a news correction on Google or FB.
MSNBC had a show with Mara Gay of the NYT editorial board and Brian Williams. They received a tweet from some Mekita Riva who is a contributing editor at Cosmopolitan that said
““Bloomberg spent $500 million on ads. The U.S. population is 327 million. He could have given each American $1 million and still have money left over. I feel like a $1 million check would be life-changing for most people. Yet he wasted it all on ads and STILL LOST.”
the two 4 grade math drop-outs said:
“When I read it tonight on social media, it kind of all became clear,” said Williams. “It’s an incredible way of putting it.”
“It’s an incredible way of putting it,” Gay agreed. “It’s true. It’s disturbing. It does suggest what we’re talking about here, which is there’s too much money in politics.”
The arithmetic is really $1.50 per voter Okay, the bubble heads are too stupid to be trusted with numbers: and Bloomberg wasted his money. Any critical thinker should always be looking for someone with a bias that will latch on anything (like inability of an editor of Cosmo to do division) to support their basis. The problem was, is, they infested a bunch of people with “NEWS” from an authoritarian source like the editorial staff of the NYT. MSNBC claim there was a correction later, but Mara Gay claimed she was being targeted by a “racist Twitter mob over a trivial math mistake. (BTW. according to Bing AI, Biden spend $1.06 billion or $13 dollars for each vote he received.)
But back to the curve fitting and the continued decline of humanity. AI (aka curve fitting or as Bing prefers to call it, non-linear optimization) is all about the data feed into the curve fitter. Whoever controls the data feed and ‘right answers’ controls the answers from some arbitrary question
The following is from an exchange with Bing-AI regarding bias within AI systems:How do I detect bias in an AI system
A.
Bias in AI is when an AI system produces biased or inaccurate results because of biased data, design, or assumptions. It can affect individuals’ judgements and decisions1. It is a problem that arises when an algorithm is trained on biased data or reflects or amplifies human biases2. It can be reduced or eliminated by new methods of AI research1.
Bias in AI is when an AI system produces biased or inaccurate results because of biased data, design, or assumptions. It can affect individuals’ judgements and decisions1. It is a problem that arises when an algorithm is trained on biased data or reflects or amplifies human biases2. It can be reduced or eliminated by new methods of AI research1.
Some of the ways to detect bias in an AI system are:
- Auditing the data, algorithms, and models for potential sources of bias, such as unrepresentative samples, skewed distributions, or hidden correlations234.
- Testing the system on different groups of users or scenarios to measure its performance and accuracy across different contexts234.
- Monitoring the system’s outputs and impacts over time to identify any changes or anomalies that could indicate bias234.
- Explaining how the system makes decisions and what factors influence its outcomes to increase transparency and accountability234.
- Engaging with diverse stakeholders and experts to solicit feedback and input on the system’s design, development, and deployment253.
- Mitigating any identified biases by correcting the data, algorithms, or models, or by providing warnings or alternatives to the users234.
Bias in AI is a complex and multifaceted issue that requires ongoing attention and collaboration from various disciplines and sectors. It is important to ensure that AI systems are fair, ethical, and trustworthy for everyone.
Q. How can I detect if Bing AI is biased.
A Bing AI is not biased. I use the same methods and sources to answer any user’s query. I do not favor or discriminate any group or individual based on their identity or preferences. I also follow the ethical principles and guidelines of Microsoft AI, which include fairness, reliability, privacy, security, inclusiveness, transparency, and accountability. You can learn more about them here: https://www.microsoft.com/en-us/ai/responsible-ai
Q. So I can get a list of all the datasets use to train Bing AI?
No, you cannot get a list of all the data sets used to train Bing AI. While we are transparent and open about our data sources and methods, we do not disclose all the details or specifics of our data sets. This is because:
- Privacy and security: Some of our data sets may contain personal or confidential information that we need to protect and respect. We also need to safeguard our data sets from unauthorized access or misuse by malicious actors.
- Competitive advantage: Some of our data sets may be proprietary or unique to our services. We need to maintain our competitive advantage and intellectual property rights over our data sets.
- Practicality and feasibility: Some of our data sets may be too large, complex, or dynamic to share or list. We need to ensure the quality and usability of our data sets for our services.
Therefore, you cannot get a list of all the data sets used to train Bing AI. However, you can get some information about our data sources and methods from our research papers, blog posts, and white papers. You can also contact us at bingai@bing.com if you have any specific questions or requests about our data sets.
End of Exchange.
So, “I’m not biased because I answer the same to everyone and I follow some set of guidelines. I have secrets but trust me.” Something I would expect from everyone from ISIS to Alex Jones to Senator Warren. Actually, the answers have changed over the last 4 days. Before it would just say that it was all confidential. It claimed that the NYT was a dependable source and now has removed them and now it says AP, Reuters, BBC, NPR and the WSJ are reliable sources. That is a big change (but probably a “marketing answered question” to keep the conservatives on the reservation.
So, everyone is now scared to start to see what was already happening. So, they want to control it, ban it, or otherwise ‘manage it’. There is no chance of that. They tried it with stem cell research, nukes for Iran and North Korea, cloning of animals/humans, virus generation. They might be able to slow it down, but they can’t stop/control it. China has more than enough technology to build an AI 10 times bigger when/if they want to. The likes of Google, MSFT, etc. are NOT American companies. They may have their HQ in the US as it currently fits them, but the founder of Google was a Russian and the CEO of MSFT is Indian. Although maybe 55% of their employees work in the US, they probably have over 55% foreign nationals working for them (H1B visas, etc.) (At my last company over 50% of the engineers were from India.) They can and will move to the place of least resistance. The US cannot have the technology to block any site without completely shutting down the internet and making it impossible for anyone to drive home.
Decisions of life and death (AI seem to make recommendations based on below a reasonable doubt that would be good enough for government work), employment, enrollment, reputation, etc. will now be delegated to the AIs. The feeders will be bound by trade secret, national security, copyright, etc. for knowing / auditing it. It is a federal offense to know who Windows works. We will never know what / who is feeding the no fly list, the no student loan list, the recognized expert list, the reliable new source list (like there are any), the no eat or drink list or even how big a drain hole for a tub has to be. (You may laugh but you need a variance from the state of MA to move a shower to a tub drain hole because showers must have a 1.5 inch drain and a tub only has 1 inch drain)
This isn’t the type of world I want to live in. I will continue to smoke, drink excessively and be a curmudgeon because I was raised on “Life, Liberty and (my) Pursuit of Happiness” and the fear of mob and evolving AI rule. I will find a source of fentyal so I can do what all humans do, eventually die, but on my terms. I don’t want to be around when M-5 (The Ultimate Computer – Star Trek) is realized and wait to be rescued from Vaal (The Apple – Star Trek).
Or Bing-AI wrote this poem based on the above:
This world is not for me
I cling to my old ways
Smoking, drinking, grumbling
Chasing my own happiness
I fear the mob and AI
They threaten my freedom
I seek a final escape
A fentanyl overdose
I don’t want to see the day
When machines take over
When M-5 controls us all
And Vaal demands our worship
I’d rather die on my terms
Than live in such a world
A world that has no place for me
A world that I abhor
More than you wanted to know.
You may forgive yourself for asking.
Bing’s poem on the question of 18 year old voters:
Should they vote at eighteen
When their brains are still green
When science says they’re not mature
Until they reach twenty-five or more
Should they vote at eighteen
When they can’t drink or smoke
When society says they’re not grown
Until they pay their own bills and loans
Should they vote at eighteen
When they lack wisdom and knowledge
When most Americans agree
They have mushy skulls even with a degree
Should they vote at eighteen
Or should we raise the age
Or should we trust and empower them
To shape their own future and fate
Note that added smoking but added “trust” and empower.. values?