- AI For All
- Posts
- The AI Race Escalates: GPT-4o and Google I/O
The AI Race Escalates: GPT-4o and Google I/O
PLUS: Google Search Changes Forever
Hello readers,
Welcome to another edition of This Week in the Future! OpenAI unveiled GPT-4o and made it free, with many proclaiming that the movie Her is now a reality. Plus, Google stepped up to the plate and announced Project Astra (and a ton of other stuff) at this year’s Google I/O. The AI wars have truly escalated.
Let’s get into it!
GPT-4o the Humanity
OpenAI has released GPT-4o, an upgrade to GPT-4 that has an impressively emotive, human-like voice. You can talk to GPT-4o, and it responds with zero latency. Many have already compared it to Samantha, the AI assistant from the 2013 film Her, which we reviewed in our article titled The Best (and Worst) AI Movies.
Furthermore, GPT-4o is natively multimodal (like Google’s Gemini) and will be available to all ChatGPT users including free users. The only difference between account tiers is the amount of messages you can send in a given time window.
Capabilities
GPT-4o is slightly smarter than GPT-4 Turbo while being 2x faster, 50% less expensive, and less rate limited. The main differences are the native multimodality and the real-time conversational capabilities. In fact, GPT-4o can detect when you interrupt it and promptly stop speaking as a result. It can also:
You can view all of the GPT-4o demos here. There will also be a ChatGPT desktop app where you can share your screen with GPT-4o to get assistance with anything.
Why This Matters
While it’s not GPT-5, it is a significant step towards human-like AI agents, which brings with it plenty of promise and plenty of risk. This development coincides with Chief Scientist Ilya Sutskever, who led the charge on AI safety, leaving OpenAI. His colleague on the Superalignment team, Jan Leike, also resigned. GPT-4o has received mixed reactions due to its implications for the trajectory of AI, with some voicing concerns about so-called pseudanthropy given GPT-4o’s emotive voice. The arrival of zero-latency, emotive conversational AI poses interesting questions about software design itself.
Everything AI at Google I/O
If Google Cloud Next was a tsunami of AI news, then Google I/O was … two tsunamis? Google should be concerned that there’s a spy amongst their ranks because OpenAI’s GPT-4o (revealed a day earlier) is essentially the same as Google’s announcement of Project Astra, a conversational AI assistant built on Gemini that can understand scenes from live video. The latency is slightly worse than GPT-4o, but impressive nonetheless.
We can’t cover every announcement, but here are the highlights:
Gemini 1.5 Pro is being brought to Gemini Advanced subscribers and will now support a 2 million token context window for API users and cloud customers.
Gemini 1.5 Flash is a new lightweight model for speed and efficiency.
Gemma 2 and PaliGemma are new open source models.
Gemini in Workspace is getting upgraded + AI Teammate.
Android OS is being fully augmented with Gemini.
Trillium is the sixth generation of Google Cloud TPU.
Google Search
Google is rolling out its AI search features. Google’s AI Overviews will now hover above the humble blue links (which you can still filter for). According to Gartner, search volume will drop 25% by 2026 thanks to AI. It is feared that for ad-dependent websites that purely provide information, this could be disastrous and spell the end of SEO. Webmasters will have to hope that the AI gods choose to cite their website. It remains to be seen how Google will square AI with its ads business. Perplexity has ideas.
Our Take
If you’re curious how many times the word ‘AI’ was uttered during the event, CEO Sundar Pichai kept count. Google may have finally found its stride, which is bad news for OpenAI, which is trying to counter Google’s advantage of a massive user base by making GPT-4o free. OpenAI is also rumored to be working on a search product and could partner with Apple to integrate GPT-4o natively on iPhones. But what really escalates the AI wars is the almost equal capabilities of GPT-4o and Project Astra. Google and OpenAI are truly in a race to build seamless AI agents.
🔥 Rapid Fire
OpenAI and Reddit partner to bring Reddit content to ChatGPT
Technology Innovation Institute releases Falcon 2 LLM
Firebase introduces Firebase Genkit for building AI apps
Cohere introduces Command-R model fine-tuning
Microsoft invests $4 billion in France to accelerate AI
Ampere and Qualcomm partner on low-power AI inferencing
Airtel and Google collaborate on AI and cloud for India
Tech Mahindra and IBM collaborate on enterprise AI adoption
NVIDIA to build new sovereign AI infrastructure for Japan
WE publishes research on AI adoption in new tech report
GitHub publishes research on Copilot impact on developers
Hugging Face offers free GPUs to encourage AI innovation
Verizon unveils new AI tools for customer experience
Anthropic’s Claude is now available in Europe (so is Grok)
Instagram co-founder Mike Krieger joins Anthropic as CPO
Microsoft offers to relocate China-based AI employees
US Army to issue new policy guidance on use of LLMs
Senators push for $32 billion annual spending on AI
Senate passes three bills to protect elections from AI
Sony Music warns over 700 AI companies about copyright
Hire a world class AI team
Engineers who understand AI are expensive and difficult to find, and it can be hard to figure out who to trust. On top of that, 85% of all AI projects fail.
But AE Studio succeeds.
We listen to your business challenge and help you craft and implement the optimal AI solution with our team of world class AI experts from Harvard, Stanford and Princeton.
Our development, design, and data science teams work closely with founders and executives to create custom software and AI solutions that get the job done. The secret to our success is treating your project as if it were our own startup.
📖 What We’re Reading
“In its relatively short time as a mainstream technology, generative artificial intelligence (GenAI) has found a way to gain attention in every industry across the globe. While some uses of GenAI are more immediately useful than others, widespread adoption seems to be a certainty over time. The product lifecycle management (PLM) space is no exception.”
“For information services providers, generative AI represents an unprecedented opportunity to drive sustainable advantage. By scaling solutions that leverage their vast data and content assets, companies can help their customers harness the power of GenAI while generating new revenue for themselves.”