Is The Cost of Training ChatGPT Sinking OpenAI?

Today’s Motive:

• 🤑 The cost of training ChatGPT. Is it too much?

• 🔍️ OpenAI sets their sights on Google with SearchGPT.

• 💌 Sam Altman writes a letter to the Washington Post.

AI News

🤑 Is the cost of training ChatGPT too much for OpenAI to handle?

Millions of users initially rushed to sign up for ChatGPT when it first debuted. Giving off vibes of AI being the next big cash cow, but what does it cost to run a place like OpenAI and train these massive LLMs? For further reading, checkout Sebastian Moss’s article here. 

• The company is set to spend nearly 4 billion this year using Microsoft servers to run workloads for ChatGPT. Sources have stated that OpenAI's software is currently being run at full capacity.

• Investors currently value OpenAI at 80 billion and did hit the 2 billion revenue milestone back in December according to Reuters.

• ChatGPT has been seeing its user numbers decline. A trend that started last year. Some blame this on API cannibalization and developers using OpenAI LLMs for their own programs, but... there is a lot of competition out there, and some of it is open-source and free. Meta's Llama for instance.

• Other factors have forced users to switch platforms because of whistleblowers discussing safety and company ethics. There is also the lack of consistency as the LLM talks nonsense at times, and can’t properly form images.

• The cost of running OpenAI is heavily growing and monthly users are declining. Some experts even warn the company could be in serious financial trouble by 2025.

AI Tools

🔍️ OpenAI wants to take on Google with SearchGPT

OpenAI wants to shake things up with its prototype SearchGPT. A more conversational approach to searching the web. SearchGPT is currently a prototype and only available to a select few for testing.

Full article on OpenAI blog.

• Promising to bring a more conversational and relevant search style to the web, SearchGPT aims to change the way people find information.

• Search results are provided with real-time information containing more visuals, including pictures and videos within the answers.

• You will be able to ask follow-up questions for more detailed answers.

• OpenAI states that they do not want to hinder content creators from showcasing their work by showing direct sources to information.

• Big publishing names have already jumped on board with OpenAI to be the ones their LLM trains on. News Corp, The Atlantic, Vox Media, and World Association of Newspapers and News Publishers, to name a few.

The Motive:

This raises numerous questions about how information is disseminated online. Many small-time creators produce excellent content that remains unseen, largely because major domain names overshadow smaller ones. Google often prioritizes large websites over direct information. Will SearchGPT revolutionize this dynamic, or will it simply enable its established partners to grow even larger?

AI News

📰Sam Altman pens op-ed letter to Washington for a call to action.

 💌 Sam calls on the U.S. government to let U.S. big tech continue to lead in AI advancement. States that if the U.S. doesn't, China will.

Full article and information source here.

"We face a strategic choice about what kind of world we are going to live in: Will it be one in which the United States and allied nations advance a global AI that spreads the technology's benefits and opens access to it, or an authoritarian one, in which nations or movements that don't share our values use AI to cement and expand their power?"

• Altman argues that authoritarian governments will use advanced AI technology to hoard and control scientific, health, and educational benefits for their own power-mongering.

• Sam pressed on the need for more robust security features to ensure a coalition maintains the lead in development and safety.

• He previously proposed establishing a multinational organization akin to the International Atomic Energy Agency to oversee this revolutionary technology and mitigate potential existential risks.

• "We are witnessing the birth and evolution of a technology I believe to be as momentous as electricity or the internet," Altman wrote. "AI can be the foundation of a new industrial base it would be wise for our country to embrace."

The Motive:

Just recently, five U.S. Senators sent a letter directly to Sam Altman seeking answers, and now Sam has responded. In his reply, Sam outlines the severe repercussions of authoritarian controls over AI, which he argues would be detrimental to everyone. However, is this an attempt to entangle the government in policy-making while OpenAI proceeds as it sees fit? It should also be stated that OpenAI tech remains closed-source.