☕️ Third Time's The Charm - Introducing Iyo's Generative AI Earbuds

Iyo aims to make some noise to start off the week with its Iyo One "audio computer," as the startup says. Elsewhere, though, Elon Musk's xAI brings in billions more in funding along with teasing a supercomputer, Google has its hands full removing some overviews given by its AI Search feature, and rumors about Apple's WWDC 2024 circulate the web at an all-time high. Join us at AI Tangle as we untangle this week's happenings in AI.

Iyo pitches generative AI earbuds to succeed where Humane and Rabbit failed spectacularly

​Iyo, a spinout of the Alphabet X "moonshot factory," recently teased the Iyo One, a generative AI device in the form of Bluetooth earbuds. If "AI in a box" sounds familiar, it's because it does - the same niche where Iyo aims to find success is where Humane and Rabbit failed to do the same with the Humane Ai Pin and Rabbit R1. Company founder and CEO Jason Rugolo, however, says that its tried-and-tested form factor and experience in the market with the Vad Pro gives its words some weight.

​

What do we know of the Iyo One?

The Iyo One is the company's 2nd audio-related device, which the company plans to release this winter. Prices start at $599 for the Wi-Fi model and $699 for cellular, but outside of standard practices, there are no added subscription fees, unlike with Humane. This puts it well out of the standard consumer's price range, as the Iyo One aims to be a premium device. Rugolo believes the Iyo One will provide value from day one by focusing on the features it will ship with on top of simply being a great pair of headphones, unlike the R1 and Ai Pin, which are nothing without their AI capabilities.

​Ellipsis will review your code (and do a good job!)​

Programmers sometimes write bugs - it's just part of the job. Ellipsis is an AI developer tool that automatically reviews code and fixes bugs. On average, Ellipsis decreases time-to-merge by 13%.

Ellipsis uses a LLM to review pull requests for logical bugs, style guide violations, and other issues. You can use it to ask questions about a codebase, summarize what's changed, or even make sprawling code changes.

It's like having a second pair of eyes to prevent problems and a second pair of hands to fix them.

Try it out yourself with a free 7-day trial.

What to expect going into Apple's WWDC 2024

With WWDC 2024, Apple's World Developer Conference, set to kick off on the 10th of June, Apple may introduce practical AI features like voice memo transcription and auto-generated emoji alongside a potential OpenAI partnership. Rumors about integrating chatbots deeper into the company's operating systems have also been circulating, with OpenAI getting first dibs and Google with its Gemini potentially not much later. Other thought-to-be-showcased news includes a "smart recall" function, better Apple Watch functionality, AI-powered improvements to on-device Spotlight search, and a more naturally sounding Siri.

​

Elon Musk plans for xAI to build a giant supercomputer with Oracle

Elon Musk's AI development startup, xAI, plans to build a supercomputer by fall 2025 to enhance its AI chatbot Grok, said Elon Musk in a recent May presentation to investors, according to an article from The Information. The proposed partnership with Oracle could create a GPU cluster four times larger than current ones, utilizing Nvidia's H100 GPUs. Musk claimed that if training Grok 2 took 20,000 of Nvidia's H100 GPUs, Grok 3 would require about 100,000 of them.

​

Google hurries to remove weird AI Search answers by hand

Following the fiasco of Google's latest AI Search feature recommending users to put glue in pizza and eat plain rocks, the company is scrambling to deal with the fallout. People like Gary Marcus, an expert on AI and an emeritus professor of neural science at New York University, and Meta's AI chief, Yann LeCun, agree that going from something like 80% accuracy to 100% accuracy requires much more sophisticated AI with even more complex thinking - something perhaps not doable without an artificial general intelligence (AGI).

​

Ex-OpenAI employees say that AI companies can't be trusted to govern themselves

Former OpenAI board members Helen Toner and Tasha McCauley argue that AI companies, including OpenAI, cannot self-govern effectively, according to an ​op-ed for The Economist​. They cite the reinstatement of CEO Sam Altman, despite allegations of a toxic work environment, as a failure of self-governance. The two of them say that government intervention is needed to create "effective regulatory frameworks," warning that these must be independent of AI companies to "avoid regulatory capture."

​

Elon Musk's xAI gets a surge of $6 billion in latest round of funding

Along with the supercomputer news, xAI has recently raised $6 billion in a Series B funding round, reaching a pre-money valuation of $18 billion according to the man himself. The funds will be used to bring its first products to market, build infrastructure, and accelerate R&D along. xAI directly competes with the likes of Google, Meta, and OpenAI, though its latest publically accessible model, Grok 1.5, hasn't been shaking up the competition too much. However, the company does have plans to expand Grok's user base beyond its own platform, X/Twitter.

​Antispace - Antispace is your digital brainstorming partner. It probes with targeted queries, aiding in problem dissection and solution formulation, then takes charge of implementation.

​Cubby - Manage and store all of your store files and links with Cubby, the private workspace where you can annotate your content and synthesize it with AI.

​Watto - Conversational AI made simple - Watto helps you build natural-sounding AI voice agents for every use case, from collecting leads to automating customer support.

​Sidekick - Sidekick is an AI-powered scheduling software that uses artificial intelligence and language processing to simplify scheduling meetings and one-up productivity.

A Sci-fi Concern Not Present at Hand (5-min read)

Y Combinator's CEO, Garry Tan, recently went on stage at The Economic Club of Washington, D.C., to talk about AI regulation, exploring the balance between innovation and control in a rapidly evolving technological landscape.

The Past & Future of AI (40-min watch)

The latest episode of the 60 Minutes Marathons has a few sections about AI, covering the godfather of AI who started it all, as well as what the AI revolution could have in store for the future.