OpenAI, Google, and Apple reignite the battle for AI supremacy. Who will be the king of the "Her" era?

OpenAI, Google, and Apple are all betting big on AI assistants and releasing a series of major updates. The author shares the latest information on related AI products, let's take a look.

Starting next week, Silicon Valley tech giants will kick off a new round of AI wars. OpenAI, Google, and Apple will all bet on AI assistants and release a series of blockbuster updates. Are you ready?

A new round of AI war is about to begin!

Next Monday, OpenAI will launch an online live broadcast to officially announce the upgrade of GPT-4, and there will even be a super "AI assistant" waiting for us.

OpenAI's "Head of Audio AGI Research" Alexis Conneau has changed his homepage background and is in sync with Ultraman - we will witness Magic next week.

OpenAI research scientist Bowen Cheng even said that this is much cooler than GPT-5.

All these things suggest that the real "Her" is about to appear.

Under pressure from OpenAI, Google will announce new model progress at the I/O conference the next day. It is rumored that it will also release a personal digital assistant called "Pixie" powered by Gemini.

Immediately afterwards, Microsoft will hold the Build developer conference on the 21st, and will most likely integrate OpenAI's latest capabilities into its own product line, and may even reveal its latest 500 billion parameter self-developed large model MAI-1.

There is also the much-anticipated Apple WWDC conference, which will release the iOS 18 system with integrated generative AI capabilities and put ChatGPT into the iPhone.

A series of blockbuster releases, one after another, left no chance for other companies to breathe.

One netizen asked, "Is Apple abandoning its own 'AJAX' artificial intelligence system and working fully with OpenAI? Or is OpenAI just a stopgap measure until their AI capabilities catch up?"

Apple insider Gurman summarized Apple's AI strategy:

– Device-side LLM (self-developed)

– Cloud LLM (self-developed)

- Chatbots (maybe OpenAI, maybe Google)

Apple doesn’t plan to build its own chatbots, but it recognizes there’s demand for them and is looking to acquire the technology from outside, a strategy similar to its approach to search.

Obviously, the current situation is that OpenAI is tied together with Microsoft and even Apple through AI cooperation, leaving Google alone.

I wonder who will win or lose in this battle for AI supremacy?

1. ChatGPT can make calls, and more information will be revealed during the live broadcast on Monday

The focus of the entire network is still on OpenAI.

The topic of "what will they release" has only become more and more popular, and few people are discussing the Google I/O conference.

Regarding Monday’s release predictions, netizen Ananay has made a new discovery:

ChatGPT may have the ability to make calls.

In fact, we can see this function from the following code, including keywords such as call and reject.

Additionally, OpenAI has deployed webRTC servers to implement this functionality, which were recently configured.

At first, netizens thought that OpenAI deployed the WebRTC server for voice-only mode, but now it seems that this is not the case.

Because this feature is provided by Livekit. (This is a solution that can provide real-time audio and video communication)

The netizen below commented, does this mean that ChatGPT can proactively call me without me initiating the call first?

He raised this question because in the movie Her, the artificial intelligence assistant Samantha took the initiative to call the male protagonist to tell him something.

Imagine how magical it would be if the ChatGPT assistant proactively called you to remind you or check your user habits.

However, Ananay said that this requires users to choose to allow this feature.

Indigo, co-founder of Hallid.ai, also made a comprehensive prediction/trend guess.

According to Indigo, the new version of GPT-4 should be divided into multiple versions according to different parameter scales.

Yesterday, some netizens speculated that there might be versions of gpt4-lite, gpt4-auto, and gpt4-lite-auto released.

The gpt2-chatbot that appeared in the LMSYS arena a few days ago may be a lightweight new version of GPT-4. Moreover, this means that the mission of GPT-3.5 is coming to an end, and the latest lightweight version may be free to use, while the API price will drop significantly.

As for the "magic" that Ultraman mentioned, it may be the upgraded GPT-4 - gpt4-auto, which has the ability to autonomously perform agent tasks, stronger memory and planning ability.

Of course, "AI Assistant" also brings Her into reality.

Source: indigo

Yesterday, OpenAI video generation research scientist Will Depue posted a logo of the advent of the singularity, perhaps hinting at something.

2. Google may launch AI assistant Pixie

At this critical moment in the competition with OpenAI and Microsoft, Google made it clear that all the content released at this conference is about AI.

According to Google's official website, this year's I/O conference will be held at 1 p.m. Eastern Time on May 14.

It is speculated that Google will integrate generative AI into its search engine to allow users to conduct conversational searches.

Google has also been testing new search features, such as AI conversation practice for English learners and the ability to generate virtual try-on images when shopping.

Not only search engines, but more Google applications will also integrate AI functions more deeply, such as helping users find suitable restaurants, shopping malls and electric vehicle charging stations in Google Maps.

What should I do if the call to customer service takes too long to be transferred?

The new AI feature being tested by Google can even help you automatically wait for the call to be transferred until someone answers the call and then notifies you.

In addition to various applications, the operating system cannot be left behind.

The developer preview of Android 15 was released last month, and Google will further introduce the new features at the I/O conference, possibly adding deeper Gemini integration.

Currently in the Android system, the function of generative AI is mainly driven by Gemini Nano and used in various software functions.

For example, Magice Compose can provide reply suggestions in applications such as Google Messages, and Cinematic Wallpaper uses machine learning to help users customize screen wallpapers.

Can you imagine what kind of more personalized user experience Android with further participation of AI will bring? For example, a smarter mobile phone home screen, lock screen interface and notification bar?

At last year’s I/O conference, we saw Gemini, a large language model that competed with ChatGPT. Will there be any new models this year?

In addition to the new version of Gemini, you can also look forward to the large image and video models launched by Google.

A netizen on Reddit revealed that Google has three models in stock that are being tested but have not yet been released to the public. They are expected to debut at the 2024 I/O conference.

The three models are the image generation model Imagen 3, and two models Juno and Miro that can optimize and complete images.

It is said that Miro will also have the function of video generation.

In addition, Google may release a new version of its AI assistant "Pixie" at this year's I/O, which may replace its existing similar product, Google Assistant.

Pixie is driven by the language model Gemini and is installed on Google's own hardware device Pixel. We don't know whether it will be open to other third-party devices.

But we probably won’t see an updated version of the Pixel product at this I/O conference. Google has recently released a new version, Pixel 8a, and it is now available for users to pre-order.

The appearance of the new version of Pixel 9 leaked online

It is expected that the Pixel 9 and the foldable Pixel 9 Pro Fold will be released this fall.

3. Apple is clinging to straws

At the same time, facing the aggressive impact of OpenAI and Google's AI voice assistant, netizens shouted to Apple:

Time is running out for Apple!

Although there are reports that OpenAI and Apple are about to finalize a cooperation agreement to enable ChatGPT to be installed on the iPhone and provide new generative AI capabilities for this year's iOS system.

But Apple is not ready to give up its own Siri.

Recently, the New York Times reported that Apple will upgrade and reorganize Siri to deal with other chatbot competitors.

This decision had already been made.

At the beginning of 2023, Apple executives Craig Federighi and John Giannandrea felt a deep sense of crisis after spending several weeks testing OpenAI's new chatbot ChatGPT, which was once the most popular.

They believe that the emergence of generative artificial intelligence makes Siri outdated and backward.

Siri, Apple's original virtual assistant that came with every iPhone when it launched it in 2011, has long been limited to fulfilling individual requests and unable to keep up with user-initiated conversations.

For example, when someone first asks about the weather in San Francisco and then says, "How's New York?" Siri often misunderstands the user's question.

But ChatGPT knows that the user wants an answer to the latter question.

After realizing that new technology had surpassed Siri, the tech giant launched its most significant restructuring in more than a decade.

Apple is determined to catch up in the tech industry’s AI race, and it has made generative AI a special internal flagship project, organizing its employees around the once-in-a-decade initiative.

1. Siri Super Evolution

According to three Apple insiders, Apple will release an improved Siri at its annual developer conference on June 10 this year.

The underlying technology in the new version includes new generative artificial intelligence that will allow Siri to chat with users rather than answering one question at a time.

And make Siri more conversational and more versatile.

The Siri update is part of Apple's lead in fully embracing generative AI.

To support its new Siri features, it also added more memory to this year's iPhones.

Apple has also discussed the possibility of partnering with several companies, including Google, Cohere, and OpenAI, to gain access to the AI ​​models that power chatbots.

On the other hand, Apple executives are also worried that emerging AI technologies will replace iOS as the main operating system in the future, threatening Apple's dominance in the global smartphone market.

Moreover, this new technology may also facilitate an ecosystem centered around AI applications (AI agents).

That could put a dent in Apple's App Store, which generates about $24 billion in sales each year.

But what Apple is more worried about is that if it fails to develop its own AI system, the iPhone may become a "dumb phone" when compared with other advanced technologies and lose the market.

The iPhone currently accounts for 85% of global smartphone profits and has generated more than $200 billion in sales. It can be expected that this loss is immeasurable and unacceptable to Apple.

The sense of urgency has prompted Apple to cancel another major investment, a $10 billion self-driving car project, and redirect hundreds of engineers to work on AI.

In addition, Apple will continue its usual consistency in device process tools and explore the creation of servers powered by iPhone and Mac processors.

According to insiders, Apple's upgrade to Siri is not to let it compete with ChatGPT in content generation such as poetry creation, but to let Siri focus on its original tasks:

This includes setting alarms, creating calendar reminders, adding items to a shopping list, and summarizing text messages.

Apple plans to tout its upgraded Siri as being more personal and cost-effective than rival artificial intelligence services.

Because Siri processes requests on the iPhone, it avoids data leakage in the cloud and the cost of cloud computing.

But Apple also faces risks with the small AI systems installed in iPhones:

Research has found that smaller AI systems may be more susceptible to hallucinations than larger systems. Siri co-founder Tom Gruber said:

“The goal with Siri was always to create a conversational interface that understands language and context, but that’s a hard problem. As technology changes, we should be able to do better. We can avoid a lot of difficulties by not trying to solve all problems with the same approach.”

Apple has many advantages in the field of artificial intelligence, including more than 2 billion devices in use worldwide and a leading semiconductor team.

They can support Apple's promotion of AI products and support AI tasks that require a lot of chips, including facial recognition.

2. Can Apple turn the tide in one month?

But over the past decade, Apple has never developed a comprehensive artificial intelligence strategy, and Siri has not received any major upgrades or improvements since its launch.

At the same time, the company's limitations as a voice assistant have also reduced the appeal of the HomePod, a smart speaker, because it cannot reliably complete simple tasks, such as responding to song play requests.

John Burkey, who founded Brighten.ai, a generative AI platform, after working on the Siri team for two years, said:

"Since its inception, the Siri team has not received the same attention and resources as other teams within Apple. Different departments within Apple are often independent of each other and information sharing is limited. But the fact is that AI needs to be integrated into products to succeed."

In addition, Apple also faces considerable resistance in recruiting and retaining leading artificial intelligence talent.

Due to Apple's confidentiality, there are few research papers published and conferences attended, which is an almost unbearable disadvantage for scientists.

In recent months, Apple has slightly adjusted its usual strategy to increase the number of artificial intelligence papers it publishes, but industry researchers still question the quality of the papers and see them as Apple's marketing hype.

But for some fledgling and ambitious researchers, joining Apple and being able to become a leading member of a project is an important reason for choosing Apple.

Although Apple has adjusted its development strategy and absorbed a lot of fresh blood.

But in this massive and dazzling battle of AI voice assistants, it remains to be seen whether Apple can reverse its disadvantage at the developer conference in June.

What will the future AI voice assistant look like and how will it affect our lives?

The answer to this question is getting closer and closer.

References:

https://x.com/ai_for_success/status/1789364452640563709

https://www.theverge.com/2024/5/11/24154219/google-io-2024-what-to-expect-where-watch-livestream-ai-android-search-gemini

https://www.nytimes.com/2024/05/10/business/apple-siri-ai-chatgpt.html

WeChat public account: New Wisdom