03/06/2025
๐๐ ๐๐จ๐ฎ๐ซ๐ง๐๐ฒ ๐ญ๐ก๐ซ๐จ๐ฎ๐ ๐ก ๐๐ข๐ฆ๐๐ฅ๐ข๐ง๐
Artificial Intelligence (AI) has evolved from a theoretical concept to a transformative force reshaping industries and daily life. What began as a dream in the 1950s is now a dynamic field, driving progress and sparking debate. Today, AI technologies such as ChatGPT, Grok, Claude & Midjourney demonstrate the remarkable progress achieved, enabling machines to generate human-like text and engage in meaningful conversations. These advancements are the result of decades of research, innovation, and perseverance
Now letโs go through the AIโs Journey through the Timeline
๐๐ฎ๐ฅ๐-๐๐๐ฌ๐๐ ๐๐ซ๐ (๐๐๐๐ โ ๐๐๐๐) โ The time when AI was a series of if, this & than statement. It was in this Era that Alan Turning proposed the turning test. This was the first time somebody proposed a test to prove that AI was smarter than humans. Fast forward to the summer of 1956, and there was a Dartmouth Summer Research Project. It was during this research project that John McCarthy, an American Scientist, coined the term โArtificial Intelligence.โ
๐๐๐๐ โ ๐๐จ๐ซ๐ฅ๐ ๐ ๐จ๐ญ ๐ข๐ญ๐ฌ ๐๐ข๐ซ๐ฌ๐ญ ๐๐ ๐๐ก๐๐ญ๐๐จ๐ญ โ๐๐ฅ๐ข๐ณ๐โ. It was like an AI psychologist & would just take what you enter & turn it into a why question
๐๐ ๐๐ข๐ง๐ญ๐๐ซ (๐๐๐๐๐ฌ ๐ญ๐จ ๐ฆ๐ข๐-๐๐๐๐๐ฌ) - A report came out, causing people to lose faith in AI, and funding also dried up
๐๐๐๐ก๐ข๐ง๐ ๐๐๐๐ซ๐ง๐ข๐ง๐ ๐๐ซ๐ ( ๐๐๐๐ โ ๐๐๐๐)
1987 - The group of researchers released the Back Propagation Algorithm. This method could take its own outputs and turnaround, and feed them back into the system & learn from its own mistakes. It would repeat this process, constantly adjusting its own weights, until it received the desired outcome.
1989 - Yann LeCun developed Le Net while working at Bell Labs. Le net was the most influential Convolutional Neural Network, recognizing and understanding what it was looking at. All modern image recognition algorithms are still using the same method
This was followed by the 2nd AI Winter โ Companies went bankrupt, Research dried up, and scepticism towards AI was increasing.
๐๐๐๐ โ ๐๐ ๐๐จ๐ฆ๐๐ฌ ๐๐๐๐ค ๐ฐ๐ข๐ญ๐ก ๐ ๐๐๐ง๐ ๐๐ฌ ๐๐๐โ๐ฌ ๐๐๐๐ฉ ๐๐ฅ๐ฎ๐ ๐๐๐๐ญ๐ฌ ๐๐๐ซ๐ซ๐ฒ ๐๐๐ฌ๐ฉ๐๐ซ๐จ๐ฏ ๐๐ญ ๐๐ก๐๐ฌ๐ฌ
1999 โ Nvidia released GEFORCE 256, the first modern GPU. Unlike the CPU, the GPU was able to do a whole bunch of tasks simultaneously
2006 โ Geoffrey Hinton creates the first Deep Belief network. This enabled AI to learn from a ton of data without a whole bunch of human labeling.
2007 โ Nvidia released CUDA โ Computer Unified Device Architecture. This enabled developers to use GPUs for general-purpose development. Now GPUs were used for absolutely anything. This was the moment when things started to ramp up
๐๐ก๐ ๐๐๐๐ฉ ๐๐๐๐ซ๐ง๐ข๐ง๐ ๐๐ซ๐ (๐๐๐๐ โ ๐๐ซ๐๐ฌ๐๐ง๐ญ)
2018 โ OpenAI released GPT-1
2019 โ OpenAI released GPT-2
Alphastar from Gamemind to master Games
2020 โ We got GPT3 from OpenAI. The first natural language model that started impressing people
2021 โ OpenAI released DALL-E 1, the first transformer-based Image Generation Model
๐๐๐๐ (๐๐๐) โ ๐ ๐ก๐ฎ๐ ๐ ๐ฆ๐จ๐ฆ๐๐ง๐ญ ๐๐จ๐ซ ๐๐ ๐๐ฌ ๐๐ฉ๐๐ง๐๐ ๐ซ๐๐ฅ๐๐๐ฌ๐๐ ๐๐ก๐๐ญ๐๐๐ ๐ญ๐จ ๐ญ๐ก๐ ๐ฐ๐จ๐ซ๐ฅ๐. ๐๐ญ ๐๐๐๐๐ฆ๐ ๐จ๐ง๐ ๐จ๐ ๐๐๐ฌ๐ญ๐๐ฌ๐ญ ๐๐๐ฌ๐ญ๐๐ฌ๐ญ-๐ ๐ซ๐จ๐ฐ๐ข๐ง๐ ๐ฉ๐ซ๐จ๐๐ฎ๐๐ญ๐ฌ ๐ข๐ง ๐ก๐ข๐ฌ๐ญ๐จ๐ซ๐ฒ & ๐๐๐ฆ๐๐ง๐ญ๐๐ ๐๐ ๐ข๐ง ๐ญ๐ก๐ ๐ฉ๐๐จ๐ฉ๐ฅ๐โ๐ฌ ๐ฆ๐ข๐ง๐.
2023 โ GPT 4 by Open AI
Meta released Lama open-source model
Google launched Gemini
2024 โ Open AI introduced SORA to the world, showing us that we can get realistic images out of AI
GPT 4O & 4O1 by Open AI
2025 โ NVIDIA releases Blackwell GPU, which further accelerates the ability to train AI models ๐ฎ
๐๐ก๐๐ญโ๐ฌ ๐๐๐ฑ๐ญ?
AI is now accelerating its own evolution. As we move forward, the question isn't what can AI do? โ it's what should we do with AI?
