AI 2023. Meet ChatGPT. - page 15

[Vitaly Murlenko]

Alexey Gureyev #:
I can't check it now, only tonight. I can only say that not every VPN can be used. It took me half a day to find the right one. I installed it as an extension on my browser

I use a landline. It usually connects me to England. But there's a choice. It's called iTop VPN. It keeps threatening me and asking for money. I don't fall for it - I use the free part of it. It works very well.

[sibirqk]

ChatGPT managed to pass an interview for a job as a programmer at Google, with a salary of more than 1M rubles per month )))))

As a result, Google received a questionnaire from ChatGPT, which meets the requirements for a third-level programmer with a salary of $183 thousand per year. (approximately 13 million rubles at the exchange rate as of 7 February 2023).

http://www.gazeta.ru/tech/news/2023/02/07/19681579.shtml

[PapaYozh]

sibirqk #:

ChatGPT managed to get an interview for a job as a programmer at Google, with a salary of more than 2M rubles per month ))))

As a result, Google received a questionnaire from ChatGPT, which meets the requirements for a third-level programmer with a salary of $183 thousand per year. (approximately 13 million rubles at the exchange rate as of 7 February 2023).

http://www.gazeta.ru/tech/news/2023/02/07/19681579.shtml

Let them hire :)))

[Aleksei Beliakov]

sibirqk #:

ChatGPT managed to get an interview for a job as a programmer at Google, with a salary of more than 2M rubles per month ))))

As a result, Google received a questionnaire from ChatGPT, which meets the requirements for a third-level programmer with a salary of $183 thousand per year. (approximately 13 million rubles at the exchange rate as of 7 February 2023).

http://www.gazeta.ru/tech/news/2023/02/07/19681579.shtml

13m per year and over 2m per month?

[sibirqk]

Aleksei Beliakov #:
13m a year and over 2m a month?

Yeah, sorry, I didn't notice, I clicked the wrong number.)

[Valeriy Yastremskiy]

Vitaly Murlenko #:

I use a landline. It usually connects me to England. But there's an option. It's called iTop VPN. It's always threatening and asking for money. I don't fall for it - I use the free part of it. It works very well.

I go through Touch VPN, where you can choose the country, so I do. But it works only once in a while now. From yesterday evening to 11 today MSC wrote come back later, we love you and are waiting for you) When answering in the dialogue almost always gives an error, I reload the page, enter again, then it works.

[Реter Konow]

We are undoubtedly seeing ChatGPT rise to the crest of the hype wave. People are impressed with the AI's ability to provide detailed, science-based answers. It's even helping someone, in some ways.....

It's worth noting:

• A polite tone.
• A neutral stance.
• Maintaining objectivity
• Avoiding bias
• No flooding or aggression.

This fact suggests a careful filtering of the training data. Apparently, OpenAI was removed from the training set:

• Trolling
• Sarcasm
• Insults
• Commonplace bickering and squabbling
• "Hate speech
• Racism, radicalism, extremism and discrimination.

A person familiar with the Western MSM knows HOW much rubbish there is. The data cleansing effort was successful, though not perfect. ChatGPT did not raise a wave of user outrage. In itself, this is an achievement. Obviously, OpenAI understood how ChatGPT should NOT be presented to the public. They took into account the old mistakes of Facebook and Microsoft, whose chatbots were toxic. They tuned the model to communicate intelligently in the style of an academic discussion. The AI behaves in an extremely "user-friendly" manner, without losing its temper or humiliating the interlocutor.

I have not been able to identify

• Emotions
• Sarcasm
• Irony
• Implication
• Arrogance or dismissiveness.

Responses are balanced. The AI skilfully avoids "dangerous" topics and does not respond to provocation. In case of a "puncture" it apologises. We can conclude that it has been tested long and thoroughly.

As a result, the public accepted the new AI with interest and enthusiasm . The first stage of commercial implementation has been passed. Microsoft has invested $10 billion in OpenAI. The contract implies leasing servers and implementing ChatGPT in Microsoft products and services.

Of course, the hype will subside, people will get used to it, and what next? Will the investment in OpenAI pay off?

To answer this question, we need to understand what is the practical use of a large language model imitating an omniscient academician, often unable to solve simple logical problems, unable to see his own contradictions and unable to use the knowledge he produces?

The answer is ambiguous. More on this in the next post.

[Реter Konow]

How useful is a language model in practice? What is the purpose of a HUGE language model like ChatGPT?

AI friend? AI girlfriend? An advanced Tamaguchi? A new search engine? Wolfram Alfa in an intelligent shell?

No, and here's why:

1. ChatGPT is an innovative, interactive language interface through which the user can super-efficiently interact with a huge base of all the knowledge accumulated by mankind.
2. A convenient tool for diverse processing of the information received.
3. An assistant in learning and working in different spheres of life.

It is important to understand, at the moment we have a basic model. Profile sharpening in all areas will begin after the release.

How will this happen?

• Thefirst stage is the integration of ChatGPT into Microsoft products and services. The model will support work in Word, Excel and PowerPoint environments: generate, edit and design texts, tables, graphs, presentations, reports and articles. Also, translate languages on the fly within Microsoft programmes. If the Whisper service (voice interface from OpenAI) is connected, the work of secretaries and students will become fantastically fast and convenient.

• The second stage is the introduction of ChatGPT into Visual Studio. AI will become a permanent assistant in the programming environment. Its niche, generation of templates, code examples, general intellectual support - accompanying the work on the project with examples from documentation, search for solutions in the codebase and external literature. In addition, an additional AI module will be trained to generate and edit the graphical interface of programmes.

• The third stage is the AI's expansion beyond the Microsoft ecosystem. Possible immediate directions:

• Gaming industry. Introduction of ChatGPT into games to enhance the dialogue capabilities of NPC characters. Implementation into gaming platforms such as Unity and Unreal Engine5... In addition to generating basic scripts when creating games, the Model will be trained to work with the "physics" of game objects and characters, fill game levels with thematic scenery (using Midjorney as an example, it will be cooler than Zuckerberg's Metaverse), generate the game interface, select object textures, add sound and visual effects. The work of game designers and developers will be simplified.

• Customer service areas. Call centres will be the first to receive ChatGPT. Old chatbots on websites will be replaced. All new ones will get a powerful intelligent upgrade from the language model.

All the uses of ChatGPT are hard to list. As an intelligent, interactive assistant, it will be in demand everywhere. And by doctors, educators, police officers..... and 123 other professions.

However. At the early stage of AI's integration into society, it will not pose a danger of displacing humans from their spheres of activity. On the contrary. People will find it easier and easier to do routine intellectual work. When will the transformation of benefit into harm happen?

It will begin at the moment of mass job loss.

But why? AI does not possess intelligence comparable to human intelligence and is unlikely to do so in the foreseeable future. However, that doesn't change anything.

By becoming a universal intellectual assistant to humans in any job, AI will master an incredibly huge number of human activity patterns. Being the "lord" of patterns and templates by nature, it will bring human activity to such a degree of optimisation that the uniqueness of human intelligence will NOT be needed! From that point on, people will start losing their jobs en masse. Not because the AI is uniquely intelligent, but because its activities are WASTE-optimised to a limit beyond which humans are no longer needed in these activities.

[sibirqk]

What will happen to the economy when artificial intelligence arrives

How efficient new technologies will transform the economy

https://www.profinance.ru/news/2023/02/06/c82v-chto-budet-s-ekonomikoj-kogda-pridet-iskusstvennyj-intellekt.html

[Aleksei Beliakov]

What if Peter is an early version of gpt chat and it was being tested here O_o.