🎢AI's master plan for artificial superintelligence

AI-HOI AInauts,

Welcome to the latest issue of your favorite newsletter! It was an exciting weekend, because we did a lot of research on OpenAI, experimented with Claude Artifacts - and were amazed by all the cool AI videos we’ve found on X. And that's what today's program looks like!

This is what we have in store for you:

• 🎢 OpenAI's master plan: The road to artificial superintelligence

• 🌐 Claude Artifacts: How to share and remix content!

• 🤩 AI video roundup: These AI clips are crazy!

Here we go!

🎢 OpenAI's master plan: The Road to Artificial Superintelligence - Opportunity or Risk?

Imagine you wake up tomorrow and your smartphone has developed an IQ of 160 overnight. That will happen - the only question is whether in 6 months or 2 years ...

An important driver behind this timeline is OpenAI. The company presented its ambitious roadmap internally a few days ago. And what we see there is nothing less than an attempt to redefine the boundaries between man and machine.

It is therefore time to put these plans under the microscope.

🚦 OpenAI's 5-step plan to superintelligence

In an internal all-hands meeting, OpenAI presented a 5-step scale that describes the path to Artificial General Intelligence (AGI). Sounds dry at first?

Wait and see, because this scale is the key to understanding where the journey is going:

1. 🟢 Chatbots(We are here!): These are our current assistants like ChatGPT, Claude, etc. Clever, but still a long way from real comprehension.

2. 🟡 Reasoners: The next level. These systems solve problems at the level of a human with a doctorate title - and without external tools. OpenAI claims to be on the verge of this breakthrough.

3. 🟠 Agents: Imagine AI assistants that can carry out complex tasks independently for days on end. Your personal digital assistant, so to speak. And of course, with a doctorate title in all disciplines.

4. 🔴 Innovators: This is where it gets really exciting (or scary, depending on your point of view). AI systems that make new discoveries, invent things on their own and achieve scientific breakthroughs.

5. Organizations: Completely AI-driven companies that could theoretically replace an entire organization.

Will these AIs develop a consciousness at some point? Some experts believe that we will soon be having heated debates about this. But that’s for another day.

In any case, OpenAI CEO Sam Altman believes that we could reach AGI before the end of this decade. But how are they going to do it? This is where "Project Strawberry" comes into play.

🍓 Project "Strawberry": OpenAI's secret future technology

According to an exclusive Reuters report, OpenAI is working on this secret technology (remember Q*?) to take the capabilities of AI models to a whole new level.

What makes Strawberry so special? It is designed to enable AI agents to:

• Roam autonomously on the Internet

• Plan ahead

• Carry out "in-depth research"

Imagine giving AI the task of solving a complex scientific problem. Instead of just processing existing information, it could:

1. Find and analyze relevant research papers

2. Simulate experiments

3. Collaborate with other (AI) systems

4. Formulate and test new hypotheses

And we're talking about multimodal agents that can inherently understand and generate text, code, image, audio and video (text, vision, voice).

But wait a minute... If the agent is so autonomous, who controls what it does and learns there? We’ll find out, but this seems to bei OpenAI's vision for the near future.

Will OpenAI stun us with a new release soon? On the one hand, a new, improved GPT mini model has appeared in the chatbot arena - and on the other hand, Chinese VPNs have been blocked from the GPT models for a few days now. China itself had already censored ChatGPT behind the Great Firewall, but now OpenAI is also pulling the plug on their end as well. Something is going on ...

Anyway, if you look at the development of use cases from GPT-3 to GPT-4, it no longer takes much imagination to make the above-mentioned use cases a reality.

🕵️Whistleblowers and regulatory authorities sound the alarm

Plot twist! While OpenAI raves about a bright future, some employees are sounding the alarm. These whistleblowers have now filed a complaint against OpenAI with the US Securities and Exchange Commission (SEC).

The accusation? Employees were prevented from disclosing potential security risks of their technology by restrictive confidentiality agreements. Employees had to sign contracts that prevented them from disclosing information to the authorities.

Security is an ongoing issue. OpenAI has also just announced a collaboration with the Los Alamos National Laboratory (which has also developed the atomic bomb) to research the safe use of AI in the laboratory environment, particularly in the life sciences. The mandate behind this comes from the White House.

And in the midst of all these developments, Microsoft - OpenAI's main investor with a 49% share - recently announced that it would give up its seat on the OpenAI board. Instead, there will be bilateral meetings. They have seen "significant progress" over the last eight months and are confident about the direction of the company.

But what is really behind this retreat? Is it a sign of trust or strategic distancing? Given the increasing attention from antitrust authorities to the links between tech giants, this is a clear sign of pre-empting regulatory concerns (not least because of the EU Competition Commission).

And Apple also wants to give up the promised board seat for this reason …

🎯 Where do we really stand? A critical assessment

Phew, part of us can hardly wait to see how these technologies could change our lives and our work. On the other hand, we have to admit: The speed and scale of these developments are also a little frightening.

The ethical questions, the security concerns, the coming social upheavals - all of this sometimes makes us uneasy in quiet moments.

But despite all our concerns, we remain cautiously optimistic. Why? Because we see that a critical discourse is taking place. Whistleblowers have the courage to expose abuses. Big players are working together to investigate the risks.

The path to AGI seems chaotic and unpredictable - but technological progress has always been that way. What we want to say is this: Yes, we should remain vigilant and question critically - but not lose sight of the fascination and opportunities that this development brings with it.

Either way, we are living in the most exciting time that is setting the course for the future.

🌐 Claude Artifacts: How to share and remix content!

Anthropic has set the bar high with Claude 3.5 Sonnet: it's the best chatbot with projects and artifacts. This new feature allows you to create extensive, independent text documents, code snippets or interactive HTML pages - and now you can also publish, share, remix and further develop them.

Interactive content and use cases

Content and ideas can easily be transformed into apps, games, quizzes, simulations, etc. - playful and interactive. We have clicked through dozens of examples, here are the 20 best.

Games like Whack a Mole, Worm Battle, Tetris, Triple Tic Tac Toe, Pacman, Rubics Cube, Fireworks, etc.

Apps like To Do Listen, AI Video Prompt Generator, Investment Calculator, SVG Pattern Builder, Image Processor, Fake Food Delivery, …

Learn how to Prompt, to Scrape, Quantum Mechanics, Flash Cards to learn AI, Climate Change, Mindmaps, …

More in the Claude Artifacts Directory or Claude Artifacts Showcase (unofficial). Click through and try it out!

P.S.: The popular chatbot Poe has also rolled out a similar feature.

🤩AI video roundup: These AI clips are crazy!

The developments in the field of AI videos over the last few weeks have been intense; we have already reported several times on Runway, Dream Machine, KLING and Sora.

To give you a quick overview of what is already possible, we have put together the best clips for you - enjoy!

That's it for today. We hope you liked it - see you next time!

Reto & Fabian from the AInauts

P.S.: Follow us on social media - that motivates us to keep going 😁!
Twitter, LinkedIn, Facebook, Insta, YouTube, TikTok

Your feedback is essential for us. We read EVERY comment and feedback, just respond to this email. Tell us what was (not) good and what is interesting for YOU.