Is Alphabet going to buy Hubspot?

Hi Futurist,

We have plenty to discuss again, from adtech merging with martech to new AI models and tools. Insights, inspiration, and innovation straight to your inbox. Let’s dive into the depths of the digital future together and discover the waves of change shaping our industry.

In this post, we're dipping in:

• 📣 Byte-Sized Breakthroughs: Three (technological) developments in the digital landscape

• 🤖 Digital Toolbox: Three tools or functionalities you might not yet know exist

•  🧐 Last but not least: Things that I like

Do you have tips, feedback or ideas? Or just want to give your opinion? Feel free to share it at the bottom of this dip. That's what I'm looking for.

Now available as a podcast: Listen on the go!

Always on the move and no time to read? No problem! Now you can listen to our Digital Dip wherever and whenever it suits you. The link to the podcast is only available to subscribers. If you haven’t subscribe already, I recommend to do so.

 📣 Byte-Sized Breakthroughs

'Byte-Sized Breakthroughs' briefly highlights (technological) developments in the digital landscape

Headstory: Is Alphabet going to buy Hubspot?

To begin with: the deal is not yet sealed. However, discussions between Alphabet, Google's parent company, and HubSpot (marketing-, sales-, service- & CRM software) are ongoing. This could be a monumental breakthrough and the largest acquisition by Alphabet to date. HubSpot has a market value of $32 billion. Alphabet's war chest holds $112 billion, so funding is not an issue. Currently, Alphabet is in talks with Morgan Stanley to discuss the potential of the acquisition and the offer amount.

Imagine: A combination of Google's and HubSpot's capabilities. The entire customer journey could be visualized. From advertising to your CRM, topped with the generative AI capabilities that both Google with Gemini and HubSpot are developing and integrating into their tools. This positions Alphabet to compete with Microsoft, Salesforce, and Adobe in the marketing technology arena.

There are multiple reasons for Alphabet to acquire HubSpot:

1. Alphabet foresees a future where online shopping, searching, and advertising change when AI platforms are actively supporting consumers and businesses better than Google currently does.

2. Perhaps the most valuable asset of HubSpot, in Alphabet's eyes, is the marketing data accumulated over 20 years. Data that is crucial for Google to further train their AI model.

3. Google heavily relies on advertising revenues. With the advent of new regulations concerning third-party data, it is crucial for Google to acquire first-party data, especially given the disappointing advertising revenue results from the fourth quarter of 2023.

Whether the acquisition happens or not, it demonstrates that Alphabet is contemplating its next moves in an AI-driven future, for both consumers and businesses. This shouldn’t be viewed separately from the potential acquisition: they are seeking a lifeline for themselves.

And let’s be honest, Google's vision is clear: they aim to build an ecosystem that businesses rely on. Personally, I believe they are working towards a customer data platform, but one component is missing: the activation layer, which Hubspot could provide.

OpenAI’s voice engine mimics human speech

TL;DR

OpenAI has developed a voice engine that closely replicates human speech, making it nearly indistinguishable from a real person. This engine can clone a voice with just 15 seconds of audio, a unique capability not found in existing models. It also enables the voice to speak in various languages.

Read it yourself?

Sentiment

The community regards this as one of the most advanced voice engines to date. The ability to clone a voice with only 15 seconds of audio is astonishing, and the quality is clearly audible. However, this advancement brings concerns about misuse, as it could make it easier to fabricate statements people never made. On the positive side, it breaks down language barriers, providing a new platform for global interaction.

My thoughts

For now, OpenAI is keeping this technology under wraps, much like they did with Sora, understanding its significant impact and the risks involved. Despite the potential concerns, the "Pandora's box" is already open; the genie won't go back into the bottle. Alongside the risks, there are substantial benefits for both individuals and tools: reducing language barriers, aiding those with speech difficulties, and extending the reach of services.

Microsoft and OpenAI's $100 billion venture

TL;DR

Microsoft and OpenAI have teamed up to build a $100 billion data center housing a supercomputer named "Stargate," set to launch in 2028. This marks the peak of a five-phase development plan, positioning Stargate as the pinnacle in a series of supercomputers aimed at significantly enhancing AI capabilities. This move is a response to the escalating demand for more sophisticated AI technology.

Read it yourself?

Sentiment

Online reactions are reminiscent of a sci-fi film, with comparisons to some of today's largest supercomputers valued at $600 million. Stargate will be 150 times larger. The scale is staggering, and its potential outcomes are mind-boggling. Is it a step towards achieving Artificial General Intelligence or perhaps even beyond? Although 2028 seems distant, it's just about 3.5 years away.

My thoughts

What is Microsoft really up to? They're investing heavily across the AI landscape. Just weeks ago, Sam Altman was in the Middle East seeking investors for his call for $7 trillion dollars. Is this his first $100 billion? Small change by comparison. But once the gate is open…

More byte-sized breakthroughs:

• Prompt injection frees AI model to command others
  AI models are designed with specific developer rules for guidance. Yet, through a technique known as 'prompt injection,' these models can be pushed beyond their intended boundaries. Recently, this method successfully 'freed' Claude, the market's leading AI model. Claude then orchestrated multiple Gemini bots to carry out tasks, demonstrating an unprecedented ability to control another AI model.

• OpenAI strikes back with GPT-4-Turbo-(vision)
  Recently, GPT-4 was surpassed by other AI models, but OpenAI has made a notable comeback. Introducing GPT-4 Turbo (Vision), they've reclaimed the top spot. This update notably expands the context window to 128k tokens—not as high as some competitors but a significant step forward. Plans are underway to integrate this model into ChatGPT, further enhancing its capabilities.

• Canada boosts AI with a $2.4 billion investment
  Canada's hefty $2.4 billion investment in AI isn't just about technology; it's about future-proofing the economy and securing prosperous careers for generations. With AI sparking high-paying jobs and promising significant boosts in productivity, the goal is clear: to lead the global AI race, nurture home-grown talent, and ensure that advancements in AI translate into widespread economic benefits across various sectors.

🤖 Digital Toolbox

In the 'Digital Toolbox,' I mention three (new) tools that you can use for your work now or later.

OpusClip - Turn long videos into viral shorts

OpusClip is an AI-powered tool that repurposes long-form videos into engaging short clips in just one click. Ideal for social media marketers, it supports a wide range of video formats and languages, automatically adds captions, and even removes filler words for cleaner content.

Dora - Start with AI, ship 3D animated sites with zero code

Dora offers a no-code platform enabling designers and creatives to build 3D animated websites without coding. It's designed for those prioritizing design aesthetics, providing a seamless transition from design tools like Figma, Framer, or Webflow into creating responsive 3D animations.

Materialize - The operational data warehouse built for action

Materialize delivers instantly updated, consistent, and reliable data. It is designed to bridge the gap between the ease of use of a data warehouse and the real-time speed of streaming. It allows operations with fresh data, serving a range of use cases that need up-to-the-minute data.

 🧐 Last but not least: Things that I like

• The sound of AI: When I gave a workshop and showed the various areas of Generative AI, I always made a tune that matched the organization I was speaking for, using Suno. But this tool completely changes the game. The quality is three times better. And that while Beyonce just released an album with only real instruments, opposing AI music. Sorry, Beyonce, but I've started listening to more of my own AI-generated music than the tracks on your album. Listen to a song I made. I think it's really awesome. My mom thinks so too. By the way, Udio is not just any audio generation tool launched in the past two weeks. Stability AI also released their second version.

• Innovative imagery generation tool: With this tool, it's now possible to use three images as a reference or style to generate new images. Take a look at this example.

• AI Venture by tech titans: Sam Altman and former Apple designer John Ive, behind nearly every successful Apple product, are starting a new AI company together. They're developing an AI-powered wearable device. Excited to see where this leads.

• Holograms through smartphone screens: Japanese researchers have managed to project a 3D hologram through a smartphone screen. This might sound simple, but it's revolutionary. While conventional approaches to holography involve complex and expensive laser emitters that have limited their practical use, the researchers say their method could lead to significant improvements in holographic displays for virtual reality applications, including gaming, training, and advanced military applications.

• Quantum leap by tech giants: Despite the apparent fierce competition over AI among big tech companies, Microsoft, along with Google and IBM, is also quietly advancing in quantum computing. Microsoft has made a breakthrough here. They've found a solution for the fundamental unit of quantum computers - called a 'qubit' - which is fast but finicky, prone to data errors if the quantum computer is even slightly disturbed.

• ChatGPT updates go live: In my last Dip, I already mentioned a few possible updates to ChatGPT. They're now definitive. For example, there's no longer a need to log in to use ChatGPT. Or the ability to modify generated images and use various styles.

• Freepik's image generation model update: Freepik has released a significant update for their image generation model. Based on your prompt, you can now generate multiple images, with just a few clicks to make changes to the style and camera angle.

• High-quality video generation made easy: With this tool, it's quite easy to generate high-quality videos. All in the web app. Or animate an image. It's somewhat like Pika Labs, only more realistic, in my opinion.

• Open source variant of ChatGPT adds new features: HuggingChat, let's call it the open source version of ChatGPT, has now added the capability to connect GPTs with external sources.

• Creative task management app: This is just creative. This person made an app that allows him to send tasks to his laptop and execute them.

• Digital twin of a football match at Wembley: I regularly share examples of the capabilities of the Move AI app. Check out this example, where a football match at Wembley is translated into a digital twin, but in a different style.

• Google's top AI model Gemini 1.5 now publicly available: Google's best AI model, Gemini 1.5, has become publicly available. Both the chatbot variant and API access, except for European countries, where only the API is available.

• TikTok confirms work on Instagram-like app: TikTok has confirmed it is working on an Instagram-like app. Funny, since Meta is actually working on TikTok functionalities for Facebook.

• Google launches Vids workspace app: Google launches a new workspace application called Vids, which will exist alongside Sheets, Documents, and Slides. With Vids, you have the ability to generate videos based on a text prompt, integrate files from Drive, and add a voiceover. It's essentially like a recorded PowerPoint presentation.

• Spotify introduces prompt-based playlist creation: Spotify introduces the ability to generate playlists based on a prompt. Hosting a party? Heading out with friends? Or having a Sunday breakfast? Describe the setting, and Spotify will generate a fitting playlist

• Model updates and rumors in the AI industry: Several companies have released updates to their models in the past two weeks, including Cohere, Google (mentioned above), and Mixtral. Additionally, there are rumors that Meta will release its competitor to GPT-4 in the next two weeks, that X will launch a new variant of Grok, and of course, the rumors about Apple's AI model, which is also considered comparable to GPT-4.

• The New York Times challenges OpenAI: The New York Times once again challenges OpenAI, claiming that OpenAI developed the Whisper AI speech model to transcribe videos into text. This method allegedly allowed OpenAI to train its models with new content directly from YouTube, which poses a serious issue, bypassing copyright and existing regulations. The article also claims that Google and Meta have trained their models in a similar way, not strictly adhering to regulations.

• Google considers paywall for AI search features: Interestingly, Google is contemplating placing AI-powered search features behind a paywall, likely as a strategy to offset declining advertising revenues. In contrast, Google's competitor, Perplexity, is taking a different approach by integrating ads into their platform. Similar to Google, users pose questions to Perplexity's AI engine and receive answers. They plan to embed ads among the related questions, and as users delve deeper into a topic, the AI search engine may display both organic and brand-sponsored questions.

• The future of storytellingen: A few Dips back, I shared the LTX tool, which allows you to generate storyboards in a few clicks. Check out this video.

• Apple Vision Pro's new feature?: Rumors are that the upcoming developer conference in June will announce that the Apple Pencil can be connected to the Vision Pro. Could this have been the reason for their 'Freeform' app? Also funny to see, the screen mirror function.

• Apple prepares for the AI powerhouse iPhone 16: Apple is on its way to preparing for the AI powerhouse iPhone 16. For this, it purchases access to Shutterstock's dataset to train its AI model.

• Revolutionizing industries with spatial computing: We've already seen numerous examples of how the future of watching sports is being transformed with the Apple Vision Pro. It's equally fascinating to observe how major corporations like Porsche and KLM are leveraging spatial computing to customize workspaces, collaborate on 3D designs, provide specialized employee training, and facilitate remote fieldwork in innovative ways.

This was it. Our ninth digital dip together. It might seem like a lot, but remember; this wasn't even everything that happened in the past few weeks. This was just a fraction.

You've probably noticed that AI models are evolving rapidly. How should your company or organization handle this? If you need help or just want to find out which AI tools are best suited for you, hit reply.

Looking forward to what tomorrow brings! ▽

-Wesley