🤕AI's $600B Challenge

Google researchers raise concerns over generative AI's impact on the internet

Key Points 

• The misuse of generative AI includes altering appearances and falsifying evidence to influence public opinion, facilitate scams, or generate profits.

• Despite its findings, the paper doesn't address Google's significant mishaps with AI.

☕News - Google researchers have published a new paper cautioning that generative AI is flooding large parts of the internet with fake content. This is pretty ironic because Google itself has been actively promoting this technology to its vast user base.

Anyways, according to the study, most people using generative AI are using it to create fake or altered content such as photos or videos, which makes it harder to tell what's real and what's not online.

📓What's more? The researchers also looked at previous studies and about 200 news articles about how people misuse generative AI. They found that changing how people look and faking evidence are common tactics used in real situations. These tricks often aim to sway public opinion, help scams, or make money. 

Adding to the issue, generative AI systems are getting better and easier to use, even for people with minimal tech skills. This is distorting people's understanding of social and political realities, as well as scientific consensus. 

Even more concerning, due to the abundance of fake AI content, researchers observed cases where influential individuals were able to dismiss unfavorable evidence as AI-generated. However, in reading the paper, it's clear that what's often labeled as "misuse" of generative AI actually seems like the technology functioning as designed. People are exploiting generative AI to create a lot of fake content because it excels at this task, thus saturating the internet with AI-generated material.

🤡What's the most interesting point here? The paper doesn't seem to mention Google's own significant mistakes with this tech, even though as one of the biggest companies, they've had some really huge mess-ups. 

In any case, as companies like Google continue to integrate AI into everything, we'll probably see more of these issues cropping up in the future.

Sequoia report warns AI sector must generate $600 billion annually to sustain operations

Key Points 

• Despite heavy investments, big tech isn't seeing expected revenue growth from AI, highlighting a gap in end-user value.

• David Cahn, an analyst at Sequoia Capital, says AI companies need to make around $600 billion each year just to cover their infrastructure costs, like data centers.

• Cahn advises recalibrating profit expectations, focusing on sustained innovation and value creation to avoid economic risks from speculative bubbles.

🏭News - Despite pouring a lot of money into AI infrastructure, big tech companies aren't seeing the expected revenue growth from AI. This means there's a big gap in providing value to end-users. David Cahn, an analyst at Sequoia Capital, says AI companies need to make around $600 billion each year just to cover their infrastructure costs, like data centers.

🙃Here's why - Take Nvidia, for example, which earned $47.5 billion last year from its AI GPUs. But these GPUs are only half the cost of running a data center. The other half includes things like energy, buildings, and backup generators. So, if you double Nvidia’s earnings, that’s about $95 billion to cover all those costs.

Then, the companies using this AI infrastructure, like startups and cloud providers, need to make a profit too. To factor in their profit margins, you double the $95 billion again, which brings the total to around $190 billion. Therefore, taking into account the massive investments and the need for profits across the board, companies need to aim for $600 billion to keep everything running and profitable.

That being said, there's another challenge to the optimistic view of investing in AI infrastructure. Unlike physical infrastructure, AI GPU computing might become cheaper as new players like AMD, Intel, and even Google, Meta, and Microsoft with their custom processors enter the market. This could lead to intense price competition and make it harder to achieve high profits.

💡The solution - David Cahn believes the industry should lower its expectations for quick profits from AI. He emphasized that recognizing the speculative nature of current investments and focusing on sustained innovation and value creation is crucial. If not, the bubble worth hundreds of billions could burst, possibly leading to a global economic crisis.