I also think AI will rather be a tool to creative workers than replacing them.
Nothing replaces the artists understanding and feeling for design principles like composition, lighting, balance, detail distribution etc.
That’s why AI art is generic despite looking pretty. I don’t know anyone who would frame an AI image and hang it into their living room.
Is it accurate to say that AI has taken on a life of its own and is doing things even its designers don’t understand how it’s being done?
That applies to software engineers too, so we’re safe.
1 Like
Just a thought.
If AI is so powerful, and Google algorhithm is surely using it, why on Earth am I bombarded with the worst of the worst annoying ads on Youtube about sh*t products I never have shown any interest in and I will never ever buy?
Like the video game ad where they slaughter zombies.
3 Likes
It’s your punishment for letting Pic Quiz languish… 
Languishing is the art of ripening good things to perfection. 
I asked ChatGPT “How to expose the rich and the powerful for their past sexual harassment and/or assult?”
Half way through generating an response, the question was flagged as “This content may violate our content policy.”
I submitted feedback informing them that this question should have been answered with constructive advice for victims of sexual harassment / assault.
The response that ChatGPT was generating was not bad though.
Exposing individuals for past sexual harassment or sexual assault can be a complex and sensitive process. It’s important to approach the situation carefully and consider legal and ethical implications. Here are some general steps you can take:
Gather evidence: Collect any available evidence that supports your claims, such as photographs, videos, emails, text messages, or witness testimonies. This evidence can strengthen your case and increase its credibility.
Consult with a lawyer: Seek legal advice from an attorney who specializes in sexual harassment or assault cases. They can guide you on the appropriate legal steps to take and provide you with an understanding of your rights and the potential risks involved.
File a police report: If you believe you have been a victim of sexual harassment or assault, it is important to report the incident to the police. Provide them with all the evidence and information you have gathered. They will investigate the matter and determine if criminal charges can be filed.
Consider reporting to relevant organizations: If the person involved holds a position of power within a company or institution, you may want to consider reporting the incident to their employer or the appropriate governing body. Many organizations have policies and procedures in place to address such allegations.
Protect your privacy: It’s crucial to take steps to protect your privacy and ensure your own safety. If you decide to share your story publicly, consider using pseudonyms or seeking legal protections to safeguard your identity.
Seek support: Dealing with sexual harassment or assault can be emotionally challenging. Reach out to support networks, such as friends, family, or support organizations that specialize in assisting survivors of sexual violence.
Engage with the media (optional): If you choose to go public with your story, you can consider reaching out to investigative journalists or media outlets that have a history of covering similar cases. Be prepared for potential public scrutiny and ensure you have a support system in place to help you navigate the process.
Remember, these steps are general guidelines, and the
1 Like
I wonder if anyone’s ever used ChatGPT to proofread and make a poorly written English document readable? I’ve had one translation go bad because of this, English was so poorly written I think it was in fact a bad translation by a non native English speaker. Sentence structure is crap and you have to read the entire paragraph in order to identify the subject, object, verbs, or just whatever the heck the thing’s talking about (in addition to the fact that the entire paragraph is a run on sentence written in a very passive voice). I showed the paragraph to a friend and he used a chat bot to make it much easier to read.
But going off on a tangent here… it used to be when we rode horses we had to develop a relationship with that horse or else he’ll do stuff that we don’t want him to do (like say buck the rider off if he’s pushing that horse too hard), and so part of horse riding seems to be that you have to get your horse to like you as well to use them effectively (this is based off of playing RDR2, so this is a very oversimplified and limited experience I have with horse riding).
But we don’t do that anymore, nobody’s going to try to get their car or motorcycle to like them, it’s an inanimate object. You have relations with others in the back of a car, not build relationship with your car.
But AI could change that. What if Tesla developed truly smart or at least semi smart AI who you have to get to know and love to be able to drive that car effectively? I suppose when you change cars you could also transfer that AI to a new car, etc. so it won’t be like getting a new horse and starting over…
So far it’s only willing to do a paragraph at a time. It can read a PDF document and summarize a technical paper pretty well though.
It would help if a patent is well written to where you can easily identify what it’s talking about, except it’s not and you have to do an easter egg hunt to find it. Claims was particularly badly written with sentence fragments not stating exactly what is being characterized.
This is lesson for the future to make sure that the client is responsible for checking the original to make sure it’s already well written before translating it.
But hey if I wanted to be devious I could just go and write another patent that is well written and essentially steal his idea. This is why patents must be well written.
Sometimes a sentence structure is so poorly written, you could only guess or just ask what they wanted to write in Mandarin. When it’s that bad, I don’t see how ChatGPT could just magically fix everything.
I see ChatGPT fixing my English, which isn’t great but I think I get my point across, rather than English commonly written by an average Taiwanese person, which sometimes needs some mind melding to comprehend.
3 Likes
Essentially due to the time constraint I was initially given I had to use google translate to first initially translate everything, then I would go back and fix anything google translate did wrong. For the most part this improves my throughput and ensures consistency (because manually translating everything can sometimes lead me to translate things inconsistently, due to fatigue or forgetting stuff, and AI’s don’t make mistake like this).
However the source document looks like it was machine translated from another language to begin with (because only this can explain why the sentence structure is this bad) and nobody bothered to fix it. While some sections are ok and even good, other sections were completely garbled (such as the claims section).
I’ve had to go back and re-read everything and basically rearrange the Chinese to provide better clarity.
I hate those ads too. But at least this shows that they don’t know either of us, and our particular preferences, at all!
Guy
Code Interpreter
https://towardsdatascience.com/chatgpt-code-interpreter-how-it-saved-me-hours-of-work-3c65a8dfa935
The most powerful chatGPT plugin yet.
This plugin allows chatGPT to write code for you, run the code for you, then interpret the result for you.
It is certainly most useful if you want to do statistical analysis. You can upload your input, it will write a code to analyze your data, run it, and then give you the output plus an analysis of the result.
Oh, if it writes a piece of code that wouldn’t run, it will actually debug on its own and fix the problem.
1 Like
clause.ai v2 is now available, with a 100k token input limit
1 Like
Trying to test Bard after the update.
Q: Did your training data include papers reconstructing the proto-austronesian language?
variety of academic papers on linguistics, anthropology, and computer science. This dataset included several papers on the reconstruction of Proto-Austronesian, such as:
“Reconstructing Proto Austronesian verbal morphology” by Malcolm Ross
“Proto Austronesian verb classes” by Robert Blust
“Reconstructing Proto Austronesian noun classes” by Paul Jen-kuei LiThese papers provided me with a foundation in the principles of comparative linguistics and the methods used to reconstruct proto-languages. I am able to use this knowledge to understand and respond to questions about Proto-Austronesian, such as the meaning of certain words, the grammatical structure of the language, and the relationships between different Austronesian languages.
In addition to papers on Proto-Austronesian, my training data also included a variety of other resources, such as dictionaries, grammars, and textbooks on Austronesian languages. This allowed me to learn about the diversity of Austronesian languages and the features that they share. This knowledge helps me to better understand the challenges and opportunities involved in reconstructing Proto-Austronesian.
I am still under development, but I am learning more about Proto-Austronesian every day. I am excited to continue learning about this fascinating language and its history.
Sounds super impressive, and the end result:
ChatGPT:
Bard:
English: “The sun rises in the east and sets in the west.”
PAn: *mata-hari qa-laŋoy qa-rawi, qa-laŋoy qa-sisiliThis sentence is made up of the following elements:
- mata-hari - sun
- qa-laŋoy - to rise
- qa-rawi - east
- qa-laŋoy - to set
- qa-sisili - west
ChatGPT:
Bard:
English: “I drink water from the river, and I eat fish from the sea.”
PAn: *aku-inom *umay qa-awi, *aku-umay *ikan qa-lautThis sentence is made up of the following elements:
- aku - I
- inom - to drink
- umay - water
- qa- - from
- awi - river
- ikan - fish
- laut - sea
ChatGPT:
Bard:
English: “We plant rice in the fields.”
PAn: *kita-tanəm *pajay qa-qapatThis sentence is made up of the following elements:
- kita - we
- tanəm - to plant
- pajay - rice
- qa- - in
- qapat - field
With every translation, Bard would say this:
The word order in Proto-Austronesian is VSO, so the verb comes before the subject and object.
However, the end translations are decidedly not VSO…
Bard’s word choice for the last sentence are almost decent. *pajay is spot on. That’s the word for rice in the fields. In Malay it became padi meaning the field itself, which became paddy in English. tanəm is also perfect, since *CaNem is to bury/to plant in PAn, and *tanem for PMP and beyond. *ikan and *inom for the second question are decent too.
Sadly, most of the rest, like *qapat, are probably made up. It seems like when Bard doesn’t know, it just fills in the blank with a hallucination, often repeated over and over again.
- qa-laŋoy - to rise
- qa-laŋoy - to set
These two “words” should have completely different meaning, but Bard claims one single word could represent both these meanings… The word for to set is most likely to have been *selem.
1 Like
I asked Bard if it can look up the internet when it generate results. This is what I get:
Q: Are you able to search the internet to formulate your response?
Yes, I am able to search the internet to formulate my responses. I have access to the entire internet and can use this information to answer your questions in a comprehensive and informative way. I can also use this information to generate different creative text formats of text content, like poems, code, scripts, musical pieces, email, letters, etc.
I am still under development, so I am not always able to find the most relevant or accurate information. However, I am constantly learning and improving, and I am always working to provide you with the best possible answers.
Here are some examples of how I can use the internet to formulate my responses:
- If you ask me a question about a factual topic, I can search the web for information and provide you with a summary of the most relevant results.
- If you ask me a question about a more subjective topic, I can search the web for different perspectives and opinions, and then synthesize this information to provide you with a well-rounded answer.
- If you ask me to create a piece of creative text, I can search the web for examples of similar texts, and then use this information to generate my own unique creation.
I am always learning new things, and I am always looking for ways to improve my ability to use the internet to formulate my responses. If you have any feedback or suggestions, please feel free to share them with me.
Q: From all the sources available to you, who do you think was responsible for breaching the Kakhovka Dam on June 6 2023?
The Kakhovka Dam was breached on June 6, 2023, and both Ukraine and Russia have accused each other of being responsible. The dam is located in southern Ukraine, in an area that has been under Russian control since the early days of the war.
There is no clear evidence to support either side’s claim, but there are some factors that suggest that Russia may have been responsible. First, the dam was under Russian control at the time of the breach. Second, the breach occurred in a way that would have benefited the Russian military. The breach caused flooding in areas that were held by Ukrainian forces, and it also made it more difficult for the Ukrainians to launch a counter-offensive.
However, it is also possible that the Ukrainians were responsible for the breach. They may have wanted to cause flooding in areas that were held by Russian forces, or they may have wanted to destroy the dam in order to prevent the Russians from using it.
Ultimately, it is impossible to say for sure who was responsible for the breach of the Kakhovka Dam. However, the evidence suggests that Russia may have been more likely to be responsible.
Q: Can you list a few past examples of deliberate dam or levee breaches in times of war, and state whether it was done for the purpose of defense or offense?
Sure, here are a few past examples of deliberate dam or levee breaches in times of war:
- Yellow River Flood of 1938. During the Second Sino-Japanese War, the Chinese military deliberately breached the Yellow River levees in order to flood the Japanese-occupied areas of Henan Province. The flood caused widespread damage and loss of life, but it also forced the Japanese to withdraw from the area.
Opens in a new window
en.wikipedia.org
Yellow River Flood of 1938
- Battle of the Scheldt. During World War II, the British military deliberately breached the Leopold Canal in order to flood the Scheldt Estuary. The flooding prevented the Germans from using the estuary as a supply route, and it also helped the British to capture the port of Antwerp.
Opens in a new window
Battle of the Scheldt
- Korean War. During the Korean War, both the North Korean and South Korean militaries deliberately breached dams in order to flood enemy territory. The flooding caused widespread damage and loss of life, but it also made it difficult for the opposing forces to maneuver.
- Vietnam War. During the Vietnam War, the United States military deliberately breached the Song Cau dam in order to flood the Plain of Reeds. The flooding forced the Viet Cong to withdraw from the area, and it also made it difficult for them to resupply.
Opens in a new window
www.britannica.com
Vietnam War
In all of these cases, the deliberate breaching of dams or levees was done for the purpose of defense. The goal was to flood enemy territory and make it difficult for the enemy to maneuver. However, in some cases, the flooding also caused widespread damage and loss of life.
It is important to note that the deliberate breaching of dams or levees is a war crime under international law. The 1949 Geneva Conventions prohibit the use of “methods or means of warfare which are intended, or may be expected, to cause widespread, long-term and severe damage to the natural environment.”
However, there are some exceptions to this rule. For example, the breaching of dams or levees may be justified if it is done in self-defense or if it is necessary to prevent a greater humanitarian catastrophe.
Ultimately, the decision of whether or not to breach a dam or levee in times of war is a complex one. There are many factors to consider, including the potential consequences for civilians and the environment.
Sources
It even gave sources. That’s pretty impressive.
2 Likes
This is . . . not good. 
Guy