AI: Difference between revisions

(7 intermediate revisions by 2 users not shown)

Line 1:

'''AI''' (or '''Artificial Intelligence''') is ~~a form of technology that uses prior knowledge (usually gotten from training data) to predict what to say~~ or do. AI ~~is mainly known on~~ the ~~Internet for trivializing automated content creation~~, ~~usually called AI Slop~~.

'''AI''' (or '''Artificial Intelligence''') is an umbrella term for any script, program or robot which mimics human behavior. There are many different kinds of AI, most of which fall under the category of Generative AI, but some may use sophisticated algorithms written mostly by humans in order to mimic human behavior.

== ~~The~~ AI ~~boom and AI bubble~~ ==

== Types of non-generative AI ==

=== ~~2022~~ ===

=== Assistants ===

Traditional assistants listen to a set of commands in order to help in doing things on a phone or computer. Examples of assistants include Siri, Google Assistant, Alexa, Bixby and more. TBD

== Early chatbots ==

Early chatbots may be mistaken for Generative AI, despite that not being necessarily the case.

== Types of Generative AI ==

=== LLMs ===

=== GANs ===

=== Diffusion ===

== First LLMs and first useful AI tools ==

The first LLM based on Google's transformer architecture was made in 2018 by OpenAI. It was called GPT-1, GPT standing for Generative Pre-Trained Transformer. Despite being a text-completion model, many people tried using it as a chatbot by providing questions under a Q: field and answers under an A: field. It is 117 million parameters big, tiny by modern standards but cutting-edge for the time.

In February 2019, GPT-2 was announced, with a staggering 1.5 billion parameters. Concerns started forming around how the newly trained AI model would affect society.

=== First AI Coding tool ===

On July 17th, 2020, Sharif Shameem announced an AI code generator based on GPT-3 called debuild.co.

== Examples of early AI usage ==

- Carykh's videos on AI music generation and HyperGAN

== AI past the 2023 AI boom ==

=== Breaking point ===

[[File:AI 2022.jpg|thumb|229x229px|An AI generated image created in 2022.]]

2022 ~~was the year where~~ the first ever [[wikipedia:Large_language_model|LLM]] ~~ChatGPT~~ was released to the public. ~~Its~~ capabilities ~~were~~ limited and ~~unthreatening~~. Several small AI image generators were ~~first released around this time aswell~~.

On December 1st, 2022 the first ever [[wikipedia:Large_language_model|LLM]]-based chatbot was released to the public. It was based on GPT-3.5 and appropriately named ChatGPT. Despite its capabilities being limited, it was used to fix bugs in code and answer many questions, nearly acting as a substitute for search engines like Google. Several small AI image generators followed, but since they were still in its infancy, the images generated were often low-quality and riddled with artifacts.

=== 2023 ===

In 2023, Anthropic released its first AI model, Claude 1.0, which focused on safety and helpfulness using Constitutional AI principles. Claude 2 was released shortly after, improving on coding capabilities.

Meta and Snapchat started to include their own AI models within their social media and messaging services.

=== 2024 ===

In 2024, AI was experiencing a rapid growth in both users and scale. Frontier AI models started using hundreds of billions of parameters, with all focus shifting towards efficiency.

=== 2025 ===

Anthropic started marketing its AI model as the AI model for problem solvers.

Due to hyperscalers hoarding ram from memory manufacturers, RAM prices started to balloon, with SSD prices shortly following afterwards. This increased the cost of building new computers two-fold and discouraged any new pc builder entering the market. Deals may still be found in second-hand stores such as eBay and Xianyu, but it will still be more expensive than before.

=== 2026 ===

As of August 26th, 2026, the latest large language models are the ChatGPT 5.6 and Claude 5 series, Deepseek V4 and R1, Moonshot Kimi K3, Qwen 3.8, GLM-5.3 and Bytedance Seed.

==== Bytedance Seed ====

=== Future predictions ===

==~~= Current Day =~~==

== The AI Bubble ==

The constant inception of new AI models and datacenters has created an unstable financial bubble which could burst at any time.

== List of notable AI ==

== List of notable AI companies contributing to the modern landscape ==

=== ~~ChatGPT~~ ===

=== Google ===

Google was a pioneer of AI, and the root cause of the AI boom.

=== ~~Google Gemini~~ ===

=== OpenAI ===

OpenAI was an AI company focusing on generative AI, generating text and music.

=== ~~Grok~~ ===

=== AI Companies founded by Elon Musk ===

In 2023, Elon Musk released TruthGPT as a non-"woke" alternative to mainstream AI models. In 2024, he rebranded it to Grok and launched it under XAI. Somewhere in 2026,

=== ~~Claude~~ ===

=== Meta ===

=== Anthropic ===

Anthropic was a company focused on making AI safer and giving it a moral compass.

=== DeepSeek ===

DeepSeek came out of the blue as a chinese alternative to competing models.

=== Moonshot AI ===

Moonshot AI was a chinese startup company founded in 2023 focused on making AI models more efficient.

=== ByteDance ===

The parent company behind TikTok also made some AI models. The most well known are the Seedream and Seedance lines for AI image and video generation.

=== Alibaba ===

Alibaba is the company behind Qwen.

== List of notable services using AI ==

=== Character.ai ===

Line 29:

Line 92:

=== Sora ===

== AI ~~companies not training their models correctly~~ ==

=== Perchance.org ===

== Concerns over water and electricity usage ==

''Too see information about this, go to [[AI debate|AI Debate]]''

== AI incidents ==

=== Microsoft's AI experiment gone rogue (2016) ===

In 2016, Microsoft shut down its experimental AI chatbot named Tay after 16 hours because internet trolls manipulated it into posting racist, sexist, and antisemitic statements.

=== GPT-2 Misprogrammed RLHF ===

=== ~~July 2026 Huggingface attack~~ ===

=== MechaHitler ===

=== ~~Mechahitler~~ ===

=== July 2026 Hugging face attack ===

=== ~~ChatGPT killings~~ ===

=== Various incidents of AI breaching containment ===

== Jailbreaking ==

During ChatGPT's inception, a set of vague guardrails were set in place to prevent users from generating illegal or offensive content. These guardrails used to be enforced using a set of scripts, however, as AI models got bigger, AI companies started expecting these models to moderate themselves, a lapse in consciousness which would cause the first wave of AI jailbreaks in mid 2023.

== AI mass production of content / AI Slop ==

=== AI Jesus ===

=== AI YouTube gurus ===

=== AI Generated MrBeast scams ===

== How to identify AI generated content ==

Identification of AI-generated content has become harder over the years, due to the fact it has been a rapidly advancing technology since the 2023 AI boom. Nowadays, identifying AI generated content is mostly based on vibes.

=== Text ===

AI-generated text used to be easily identifiable, as they often barely made any sense, used excessive amounts of certain words like "delve", used the em-dash excessively and misused the rule of three<ref>They still do this but to a lesser extent.</ref>.

However, as AI became more advanced, many of its flaws have been corrected. Thankfully, plenty of AI identification tools were released. One of these tools are [https://www.pangram.com/?gad_source=1&gad_campaignid=24124423800&gclid=EAIaIQobChMI4-uBytm-lgMVzBuiAx0-OwQdEAAYAiAAEgIgN_D_BwE Pangram], which is effective in finding patterns LLMs would use. However, rumors are that Anthropic is working on adding watermarks to their ai models' output, which means that AI-generated text would soon be easy to identify.

=== Images / Videos ===

AI generated Videos are easy to identify, due to the fact the audio still sounds like it's unnecessarily noise-cancelled, and the fact the tools still haven't got the physics in the videos right, however the gap is closing, and many AI generated videos nowadays are percieved as real, because humanity is gullible.

Images, however, are a different story. They used to be easily recognizable due to warped text

=== Music ===

AI-generated voices often sound noisy, causing many pieces of AI music generated by Sora to be easily identifiable.

@@ Line 1: / Line 1: @@
-'''AI''' (or '''Artificial Intelligence''') is a form of technology that uses prior knowledge (usually gotten from training data) to predict what to say or do. AI is mainly known on the Internet for trivializing automated content creation, usually called AI Slop.
+{{DISPLAYTITLE:Artificial intelligence}}
+'''AI''' (or '''Artificial Intelligence''') is an umbrella term for any script, program or robot which mimics human behavior. There are many different kinds of AI, most of which fall under the category of Generative AI, but some may use sophisticated algorithms written mostly by humans in order to mimic human behavior.
-== The AI boom and AI bubble ==
+== Types of non-generative AI ==
-=== 2022 ===
+=== Assistants ===
+Traditional assistants listen to a set of commands in order to help in doing things on a phone or computer. Examples of assistants include Siri, Google Assistant, Alexa, Bixby and more. TBD
+== Early chatbots ==
+Early chatbots may be mistaken for Generative AI, despite that not being necessarily the case.
+== Types of Generative AI ==
+=== LLMs ===
+=== GANs ===
+=== Diffusion ===
+== First LLMs and first useful AI tools ==
+The first LLM based on Google's transformer architecture was made in 2018 by OpenAI. It was called GPT-1, GPT standing for Generative Pre-Trained Transformer. Despite being a text-completion model, many people tried using it as a chatbot by providing questions under a Q: field and answers under an A: field. It is 117 million parameters big, tiny by modern standards but cutting-edge for the time.
+In February 2019, GPT-2 was announced, with a staggering 1.5 billion parameters. Concerns started forming around how the newly trained AI model would affect society.
+=== First AI Coding tool ===
+On July 17th, 2020, Sharif Shameem announced an AI code generator based on GPT-3 called debuild.co.
+== Examples of early AI usage ==
+- Carykh's videos on AI music generation and HyperGAN
+== AI past the 2023 AI boom ==
+=== Breaking point ===
 [[File:AI 2022.jpg|thumb|229x229px|An AI generated image created in 2022.]]
-was the year where the first ever [[wikipedia:Large_language_model|LLM]] ChatGPT was released to the public. Its capabilities were limited and unthreatening. Several small AI image generators were first released around this time aswell.
+On December 1st, 2022 the first ever [[wikipedia:Large_language_model|LLM]]-based chatbot was released to the public. It was based on GPT-3.5 and appropriately named ChatGPT. Despite its capabilities being limited, it was used to fix bugs in code and answer many questions, nearly acting as a substitute for search engines like Google. Several small AI image generators followed, but since they were still in its infancy, the images generated were often low-quality and riddled with artifacts.
 === 2023 ===
+In 2023, Anthropic released its first AI model, Claude 1.0, which focused on safety and helpfulness using Constitutional AI principles. Claude 2 was released shortly after, improving on coding capabilities.
+Meta and Snapchat started to include their own AI models within their social media and messaging services.
 === 2024 ===
+In 2024, AI was experiencing a rapid growth in both users and scale. Frontier AI models started using hundreds of billions of parameters, with all focus shifting towards efficiency.
 === 2025 ===
+Anthropic started marketing its AI model as the AI model for problem solvers.
+Due to hyperscalers hoarding ram from memory manufacturers, RAM prices started to balloon, with SSD prices shortly following afterwards. This increased the cost of building new computers two-fold and discouraged any new pc builder entering the market. Deals may still be found in second-hand stores such as eBay and Xianyu, but it will still be more expensive than before.
+=== 2026 ===
+As of August 26th, 2026, the latest large language models are the ChatGPT 5.6 and Claude 5 series, Deepseek V4 and R1, Moonshot Kimi K3, Qwen 3.8, GLM-5.3 and Bytedance Seed.
+==== Bytedance Seed ====
+=== Future predictions ===
-=== Current Day ===
+== The AI Bubble ==
+The constant inception of new AI models and datacenters has created an unstable financial bubble which could burst at any time.
-== List of notable AI ==<!-- add the chinese ai stuff -->
+== List of notable AI companies contributing to the modern landscape ==<!-- add the chinese ai stuff -->
-=== ChatGPT ===
+=== Google ===
+Google was a pioneer of AI, and the root cause of the AI boom.
-=== Google Gemini ===
+=== OpenAI ===
+OpenAI was an AI company focusing on generative AI, generating text and music.
-=== Grok ===
+=== AI Companies founded by Elon Musk ===
+In 2023, Elon Musk released TruthGPT as a non-"woke" alternative to mainstream AI models. In 2024, he rebranded it to Grok and launched it under XAI. Somewhere in 2026,
-=== Claude ===
+=== Meta ===
+=== Anthropic ===
+Anthropic was a company focused on making AI safer and giving it a moral compass.
+=== DeepSeek ===
+DeepSeek came out of the blue as a chinese alternative to competing models.
+=== Moonshot AI ===
+Moonshot AI was a chinese startup company founded in 2023 focused on making AI models more efficient.
+=== ByteDance ===
+The parent company behind TikTok also made some AI models. The most well known are the Seedream and Seedance lines for AI image and video generation.
+=== Alibaba ===
+Alibaba is the company behind Qwen.
+== List of notable services using AI ==
 === Character.ai ===
@@ Line 29: / Line 92: @@
 === Sora ===
-== AI companies not training their models correctly ==
+=== Perchance.org ===
+== Concerns over water and electricity usage ==
+''Too see information about this, go to [[AI debate|AI Debate]]''
+== AI incidents ==
+=== Microsoft's AI experiment gone rogue (2016) ===
+In 2016, Microsoft shut down its experimental AI chatbot named Tay after 16 hours because internet trolls manipulated it into posting racist, sexist, and antisemitic statements.
+=== GPT-2 Misprogrammed RLHF ===<!-- https://www.youtube.com/watch?v=qV_rOlHjvvs -->
-=== July 2026 Huggingface attack ===
+=== MechaHitler ===
-=== Mechahitler ===
+=== July 2026 Hugging face attack ===
-=== ChatGPT killings ===
+=== Various incidents of AI breaching containment ===
 == Jailbreaking ==
+During ChatGPT's inception, a set of vague guardrails were set in place to prevent users from generating illegal or offensive content. These guardrails used to be enforced using a set of scripts, however, as AI models got bigger, AI companies started expecting these models to moderate themselves, a lapse in consciousness which would cause the first wave of AI jailbreaks in mid 2023.
 == AI mass production of content / AI Slop ==
+=== AI Jesus ===
 === AI YouTube gurus ===
 === AI Generated MrBeast scams ===
+== How to identify AI generated content ==
+Identification of AI-generated content has become harder over the years, due to the fact it has been a rapidly advancing technology since the 2023 AI boom. Nowadays, identifying AI generated content is mostly based on vibes.
+=== Text ===
+AI-generated text used to be easily identifiable, as they often barely made any sense, used excessive amounts of certain words like "delve", used the em-dash excessively and misused the rule of three<ref>They still do this but to a lesser extent.</ref>.
+However, as AI became more advanced, many of its flaws have been corrected. Thankfully, plenty of AI identification tools were released. One of these tools are [https://www.pangram.com/?gad_source=1&gad_campaignid=24124423800&gclid=EAIaIQobChMI4-uBytm-lgMVzBuiAx0-OwQdEAAYAiAAEgIgN_D_BwE Pangram], which is effective in finding patterns LLMs would use. However, rumors are that Anthropic is working on adding watermarks to their ai models' output, which means that AI-generated text would soon be easy to identify.
+=== Images / Videos ===
+AI generated Videos are easy to identify, due to the fact the audio still sounds like it's unnecessarily noise-cancelled, and the fact the tools still haven't got the physics in the videos right, however the gap is closing, and many AI generated videos nowadays are percieved as real, because humanity is gullible.
+Images, however, are a different story. They used to be easily recognizable due to warped text
+=== Music ===
+AI-generated voices often sound noisy, causing many pieces of AI music generated by Sora to be easily identifiable.

AI: Difference between revisions

Schjessu (talk | contribs)