AI: Weekly Summary. RTZ #997

OpenAI tests ChatGPT ads: OpenAI’s long-awaited Ads in ChatGPT are being tested now by the company. This despite its Google Gemini inspired internal ‘Code Red’ a few weeks ago. The tests come on the heels of Anthropic’s pointed and clever four Superbowl ads, promising Claude not having ads like its prime competition. OpenAI’s ads test the UI/UX approach for its initiative, while promising all types of safety and privacy guardrails for its 900 million weekly ChatGPT applyrs. The ads will be layered on various tiers of ChatGPT from free on up. Of course its prime competitor remains Google with its Gemini AI Overviews, AI Mode, and soon to come Apple version distributed via an AI enabled Siri in the coming weeks. Expect Meta and Amazon to also play their A game with AI and advertisers this year as well. More here.

Anthropic’s Big AI Data Center Plans: Anthropic is working hard this year to not just be the ‘Pepsi’ to OpenAI’s Coke. While concurrently just concluding a $30 billion additional capital raise at a $380 billion post-money valuation. This further enables Anthropic to continue ramping up its AI Data Center Compute. To the tune of hundreds of billions in investment, or multi-gigawatt AI data centers. Note that a single Gigawatt AI Data Center and Power costs $50 billion today, so the sums add up fairly quickly. These relocates are of course ahead of an expected IPO later this year. Both OpenAi and Anthropic are apparently in a race on that front this year. And of course against Elon’s SpaceX/xAI and possibly Tesla later this year as well. The hugegest headwind for them all remains ramping up appropriate Power for the data centers. Along with of course AI GPUs, memory chips, and AI research talent up and down the AI tech stack. More here.

Big Tech Stretches AI Capex Financing: Now that quarterly earnings are done, and the huge tech companies Microsoft, Google, Meta, Amazon and others are tarobtaining over $700 billion in AI capex this year alone, the time when it can all be mostly from internal cash flow is nearing an conclude. All of the huge tech companies and the up and coming AI tech companies are attempting out all sorts of equity and debt financing plans to stretch the next level of funding required. Google for example, saw strong demand for its 100 year bonds for AI financing, a rare relocate in the world of debt financing. For now investors of all stripes remain leaning in to both the funding opportunities, and have flashing green lights for the longer term AI objectives of the indusattempt. The timing on spconclude vs revenues remains the hugegest, most closely watched set of data points in the AI indusattempt. More here.

AI impact on Legal Business: Last week saw the impact of the $300 billion ‘software apocalypse’ sell-off on the broader software and SaaS (software as a service) indusattempt. All stemming from Anthropic’s relocate into legal services AI models. It’s applyful to take a closer view at both sides of that impact on the legal indusattempt that was the original catalyst. On the plus side, vertical AI native companies like Harvey are raising new funds at higher valuations. On the negative front, the actual application of AI in legal applications is displaying mixed results. Particularly since both sides in litigation are being creative in their apply of the AI technologies, which themselves are barely ready for the rigors of the tquestions and expectations. In addition, the indusattempt is wrestling with the potential impact on the core business model of the indusattempt, ‘billable hours’. Most observers agree that the net effects will take a lot more time to assess cumulatively. And this vertical will likely be a test case for other vertical industries as AI roils their business models as well. More here.

AI Fears rising again: After a short interregnum, AI Fears are rising again, particularly amongst the very AI Researchers at the top firms driving the AI scaling innovations. AI experts at OpenAI, Anthropic and others are warning of ‘rising dangers of their technology’, while some are quitting in protest. This especially as core LLM AIs obtain better and rapider, and their agent versions increasingly able to assist in building better versions of themselves. This is on top of the founder/CEOs of some of these companies voicing their own separate fears in recent esdeclares. The concerns cover the spectrum of fears from societal harm to AI driven job losses. And their voices are obtainting attention from the media and regulators at large. A lot of these fears are of course based in the very scifi that inspired many of these AI researchers to build the current technologies in the first place. Expect these debates to continue, while the core AI products are still striving to meet even the periphery of current fears and expectations AI Tech Wave. More here.