Unlocking the future of work: create your own AI agents with Zapier

Hi Futurist,

In the world of (marketing) technology, there's always something happening. It was shaken to its core following the latest announcements. 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. Happy listening. 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: Unlocking the future of work - create your own AI agents with Zapier

Anyone who knows me a little knows: Zapier is my favorite tool. When I completed my business economics degree, I wanted something completely different. It became digital marketing. Because I found data so fascinating. Zapier was my salvation in my first few months. Automating the transfer of data from A to B kept me afloat. By now, Zapier has evolved further. It's not just about moving data from A to B anymore. You can create a whole database in it or apps. And since their last announcement, it has also become your AI assistant. That’s right.

In recent months, there has been a lot of talk about AI agents. Bots that can automate tasks for you. It sometimes still seemed like future music. You definitely needed some programming knowledge for it. Let alone access to a large pool of other tools. Not anymore. In Zapier, you can now build your own AI assistant, and automate your activities even further. With the connections to 6000+ tools in Zapier, the possibilities are literally endless.

View it as an AI workplace where several AI employees of yours perform tasks for you in the background based on instructions. Here are some examples every (digital) marketer could benefit from:

1. Have it send meeting briefs to you before a meeting starts;

2. Have it send an update on the day's activities every morning, even if the tasks come from multiple different tools like Monday, Asana, or Click-up;

3. Have it write email drafts for you, so all you have to do is press send.

These are just the simple examples where you directly benefit yourself. These AI employees can also inform your colleagues about deadlines, clients about the progress of your project, or leads about potential interesting offers. And no, you don't have to be behind the controls for this. They can work while you sleep. Literally.

I wouldn't take this development lightly. Or read this and consider it a 'normal' thing. This is no longer the future. It is now. It can really save you time that you can invest in activities that you do like, that you are good at, and that energize you. This way, you become even better at the work you already do. Doesn't everyone want that?

Claude 3 surpasses GPT-4 with human IQ

TL;DR

GPT-4 has been surpassed. The British company Anthropic has launched the new version of their AI model, Claude 3. The figures are impressive. It outperforms just about every other model in benchmarks and boasts a context window of 200k tokens, achieving 99% accuracy. Astonishing. Since its release, it's been the best AI model on the market, period.

Read it yourself?

Sentiment

The reactions are overwhelmingly positive. People are blown away by the model's immense power, setting a new standard over GPT-4 for many. In fact, Claude was even put through an IQ test, where GPT-4 had previously scored a high of 85, Claude 3 achieved an IQ of 101. That's an average human IQ. Incredible.

My thoughts

The dominance of OpenAI's GPT-4 model has finally been broken. It was expected, but this leap is exceptional. An IQ score comparable to the human average raises questions about how close we are to Artificial General Intelligence. Apart from the hype, a context window of 200,000 tokens at 99% accuracy is nearly inconceivable. It means you can upload entire volumes of books, and the model will still find the needle in the haystack, a task where other models typically fail.

The EU AI Act: A forward Leap or a stumble?

TL;DR

It's official: the AI Act has passed through the European Parliament. A whopping 523 votes in favor, 46 against, and 49 abstentions highlight its acceptance. It's anticipated that the law will be published this coming May and will gradually come into effect. As the first continent to do so, EU member states now have a law based on a risk-based approach. AI systems such as social scoring or biometric identification are outright banned. Socio-economic factors like infrastructure, access to education, or employment opportunities are considered high risk. Generative AI, like ChatGPT, is deemed low risk, mainly required to comply with transparency requirements and EU copyright laws.

Read it yourself?

Sentiment

Opinions are divided. Proponents welcome the regulation for setting clear boundaries on societal applications of AI. Critics argue it's too premature, potentially stifling AI development. More critically, they question whether this unilateral move puts the EU at a competitive disadvantage globally.

My thoughts

I stand by my previous positions shared in earlier Dips. In my view, the EU's adoption of this law places it—and European businesses—at a disadvantage. Major players like OpenAI and Google restrict their most advanced AI models from European companies for a reason. This gives American companies a lead in innovation, a gap that will be hard to close. Moreover, the law likely won't be fully effective until 2026. According to Sam Altman, that's just a year shy of Artificial General Intelligence's arrival. By then, isn't the law already outdated?

More byte-sized breakthroughs:

• Science fiction becomes reality with talking AI-robots

  In our previous Dips, we've often shone a light on AI-powered robots' progress, from those handling daily tasks to ones that know when to recharge. However, a thrilling update by Figure has truly pushed the boundaries—integrating OpenAI's vocal technology allows these robots to engage in real-time conversations, reacting to immediate situations as any human would. Science fiction is becoming reality.

• Sam Altman returns to OpenAI's board
  In November, the buzz around Sam Altman's sudden exit from OpenAI, due to governance issues or a dash for Artificial General Intelligence, quickly turned into news of his return. Now, he is also back on the Board with Greg Brockman and a few notable others, including Sony's former VP, Nicole Seligman. Her involvement intriguingly hints at the fate of OpenAI's text-to-video model, Sora—will it remain within public reach? Stay tuned for more on this in the Dip.

• Elon Musk vs. Sam Altman
  Elon Musk is suing OpenAI and its founders, Altman and Brockman, for not adhering to their original mission: developing open-source AI models as a nonprofit to prevent human extinction. Musk's concern extends to OpenAI's partnership with Microsoft, fearing it undermines his vision of an all-encompassing app, given AI's potential in handling tasks that could empower Microsoft to develop a similar application.

🤖 Digital Toolbox

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

Krea - Elevate your AI images to perfection

Krea leverages advanced AI to upscale your images, ensuring they are campaign-ready with crisp clarity and stunning detail. It's the ideal tool for enhancing overall quality, making your visuals perfect for inclusion in any content.

Definite - Streamline your data, simplify your insights

Definite consolidates over 500 data sources into one intuitive platform, streamlining analytics and dashboard creation without the need for engineering or SQL knowledge. Definite turns complex data landscapes into simple, actionable insights.

Saga - Unlock the full potential of your notes

Saga redefines the way we take notes by creating interconnected knowledge networks. This tool turns your scattered thoughts, research, and insights into a structured web of information, making it easy to navigate, expand, and connect ideas, enhancing creativity and learning.

 🧐 Last but not least

• AI software engineer, Devin leads the way: Meet Devin, the first AI software engineer, ready to tackle any question you throw at him. Devin outlines steps to execute your queries, featuring an integrated code editor and browser for research, enabling complete code writing, error detection and resolution, and deployment.

• AI that sounds human: Bland AI introduces an AI speech model that sounds just like a human. Try having a conversation with it. Companies can use this model to add AI human voice in their web, app, or telephony.

• Suno v3 is a musical evolution: The text-to-music tool Suno unveils its third version. Listen to this track about Elon Musk's lawsuit against Sam Altman. Sounds almost like Whitney Houston, right?

• Pika's sound effects for generated videos: If you generate videos in Pika, they used to be silent. Not anymore. Pika introduces sound effects, automatically generating sounds for your videos.

• Upgraded ChatGPT responses: It appears ChatGPT is developing a feature to upscale output from ChatGPT-3.5 to ChatGPT-4.

• ChatGPT's Read Feature: A handy new feature from ChatGPT: now, it can read out its output to you. See how it works here.

• 3D elements on your keyboard: An inventive way to link various tools. Generate an image in Midjourney, and within minutes, it appears as a 3D figure on your keyboard.

• Midjourney describe function: The describe function of Midjourney is a handy feature to upload images and then describe them as prompts, which can be used to generate closely related images.

• Share your GPT: ChatGPT is adding new features to share your custom GPTs, for instance, with your workspace or based on an invite.

• Meta Reality Labs developed neuromotor interface: Meta has developed a neuromotor interface that translates hand movements into actions, like typing texts by moving your fingers.

• NVIDIA's local chat model: Tech giants like Apple and Microsoft are inevitably updating their operating systems' search bars to allow chatting with all files stored locally on your PC, familiar with ChatGPT. NVIDIA presents a first glimpse of this feature.

• Move AI's soccer-playing monsters: The sight is truly mesmerizing, showcasing endless possibilities for sports applications. Move AI revolutionizes human and ball motion capture, generating 3D motion data from 2D videos.

• Adobe Express mobile app: Adobe has launched the mobile app Express, allowing users to edit images on their mobile devices using Adobe Firefly capabilities.

• X releases articles: X has moved beyond being just Twitter. The era of 160 characters is over. X recently introduced a feature allowing articles to be shared on the platform.

• Microsoft Co-Pilot own GPTs: Within Microsoft Co-Pilot, it will soon be possible to create your own GPTs and share them. Co-Pilot already offers free access to GPT-4, but soon you can also create and share your GPTs within Co-Pilot.

• Salesforce prompt builder Einstein Copilot: Salesforce has added a prompt builder feature to its platform, enabling text generation and the direct use of (personal) information within Salesforce.

• OpenAI Sora available by year-end: In a discussion with the Wall Street Journal, CTO Mira Murati confirmed that Sora will be available this year, even to the general public. For now, only Hollywood has access.

• Finetune with CassetteAI: With CassetteAI, you can generate your own music. Now, it's also possible to finetune the model, allowing for the combination of different music styles to create something entirely new.

• Grok (X) going open source: Interestingly, Elon Musk announces X's AI model will soon be open source. Likely related to his lawsuit against OpenAI, accusing them of not being 'open' enough.

• Freepik's text-to-image update: Freepik has updated its text-to-image model, now allowing endless scrolling through your newly created images, a stark difference from the usual four images returned by other models.

• Midjourney's 'Character Consistency': Maintaining character consistency in Midjourney has always been challenging. That's no longer the case with the new Character Consistency function, making it easy to have generated faces reappear consistently, simplifying storyboard creation.

• Rumors of GPT 4.5-Turbo: Screenshots suggest an update from GPT-4 to GPT 4.5-Turbo. This model has a knowledge cutoff up to June 2024 and a 248k token window, double the current GPT 4-Turbo. But June 2024 is still a while away.

• Visual storytelling simplified with LTX Studio: This tool makes visual storytelling child's play. Generate videos, complete scripts, matching sounds, and even voice acting. Impressive!

This was it. Our seventh 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.

Of course, I'm always ready to provide you with even more guidance, direction, or advice on how to navigate all these developments. Or if you simply want to organize an inspiring session around all the possibilities, hit reply!

Looking forward what tomorrow brings! ▽

-Wesley