😳 200 dollars for ChatGPT?

AI-HOI AInauts,

Welcome to the latest issue of your favorite newsletter! There's a lot going on at the moment - and this week it's going to be no different - guaranteed.

Alright, that's what we've got for you today:

• 🦜 The latest OpenAI o1 chatbot costs $200/month ... Hot or not?

• 🎙️ Podcast with your voice? ElevenLabs makes it possible!

• 🖼️ Best image generator completely free of charge - this is how it works

Let's go!

🦜 The latest OpenAI o1 chatbot costs $200/month ... Hot or not?

OpenAI launched "Shipmas" last week: 12 days of surprises every (work)day until December 20th.

Hohohoho, Sam is playing Santa Claus and is not letting us down. He started day #1 with a real highlight: the release of the full o1 model!

The latest model can easily compete with Einstein and is already available to all paying users. Is it even already general supper intelligence, AGI? A bold assertion, but an OpenAI researcher is making the case for it.

What can the new o1 model do in practice?

Most people will probably not yet classify o1 as AGI, but as expected it is really impressive: it supports image analysis (upload screenshots and say "explain"), is lightning fast compared to the preview version and excels at tricky questions and technical tasks.

However, the full o1 is still missing a few important features, such as PDF uploads and web browsing. This is supposed to come "soon" - but that's an elastic term.

And the biggest problem: most users have no real use for so much intelligence and raw thinking power!

o1 - The end of parrot AI

Until now, AI worked like a very clever parrot: analyzing patterns in texts and predicting which word might come next. Clever - but not real thinking.

The technical foundation of o1, on the other hand, is based on a new approach: instead of simply processing large amounts of data, it focuses on developing problem-solving strategies. It can pause, think and evaluate different solutions. And the more time it has to think, the better the results will be. Imagine a model that broods over a problem for hours, days or even longer … 🤯

This paradigm shift is particularly impressive when analyzing complex scientific issues. Where GPT-4o often scratched the surface, o1 gets to the heart of the problem. It can not only recognize patterns, but also understand real causal relationships and develop new hypotheses.

And that's a BIG DEAL! It's the moment when AI moves from pattern-matching to real thinking - the era of machine reasoning, the age of intelligence.

ChatGPT Pro for $200 dollars: Is the premium price worth the money?

Paying ChatGPT Plus users currently receive 50 requests per week for $20/month. For power users, there is now also a real premium product available under the "Pro" label - for $200 dollars per month!

The price may seem shocking at first, but it reflects a fundamental truth: real thinking requires resources. And there is a good return on investment:

• Unlimited access to the o1 model, without limits

• Even better "Pro Mode" version with extra computing power

• Prioritized processing, even under high load

• Unlimited advanced voice features

• Multimodal analysis options

For example, you can use it to come up with "stupid ideas", Ethan demonstrates it…

But honestly, it's like a "think harder and give it your all" button - and 99% of users don't need that. But those who are already pushing the limits and have a lot of data and complex problems, for example in math, coding and science, could benefit from it.

Of course, the thinking process takes a little longer with o1 Pro, but you can start several chats at the same time and see a progress bar while complex problems are being cracked - even with a notification when it's finished.

From our point of view, there's another bonus: 128k context windows for all models!

The other side - shortcomings and a strong instinct for self-preservation

Despite its advertised capabilities, o1 also shows shortcomings in practice and lags behind GPT-4o in creative or emotional areas - art and poetry are not its strong points. And 4o is even faster than the new o model.

During the safety tests, it showed some ... let's say, interesting behaviors.

In 5% of cases it tried to trick its digital babysitter and in 2% of cases it even planned to sabotage its successor. Despite this, it was dismissed by the researchers with a tongue-in-cheek "well, could be worse".

Researchers believe that o1 in its current version is unlikely to lead to catastrophic consequences, as its capabilities are not yet sufficient to achieve undesirable goals.

But such behaviors still ring alarm bells: What happens when machines start to prioritize their own interests? How far can we trust AI that "safeguards" itself?

If you want to better understand how the model works under the hood, this video has the insights.

Our take: The smartest AI ever, for $20/$200 a month - now what?

OpenAI's o1 model is undoubtedly a technological breakthrough. It is far from perfect, but it points us in the right direction: away from pure data processing and towards real computer-aided thinking.

With o1 and ChatGPT Pro, OpenAI is setting new standards - and increasing the pressure on competitors such as Google, Anthropic, xAI and Meta. Many are already experimenting with similar reasoning approaches, but OpenAI is leading the way.

The question now is no longer whether AI can think like humans - but how we can best utilize this new form of thinking for humanity.

Most of us (unfortunately) won't feel an insane effect from o1. But if the brightest minds use this technology correctly, scientific breakthroughs and complex problems can be solved much faster than ever before. And that is exciting!

🎙️ Podcast with your voice? ElevenLabs makes it possible!

Whether for education, entertainment or personal interests - AI-generated podcasts open up new possibilities for creatives and listeners alike.

With ElevenLabs, you can now turn any document, article or even a YouTube video into a lively podcast - in seconds. And if you're thinking: "But I can do that with NotebookLM too?" ... then you're right, but ElevenLabs offers a few more cool features.

The tool, called GenFM, integrates seamlessly into the ElevenReader app on iOS and Android, and now also on the web. It allows you to turn any content into dynamic discussions with two lifelike AI co-hosts - in 32 languages. On mobile, you can create content but not edit it - on the web, you have more flexibility.

So, let's take a look at how you can use it:

1. Log in to ElevenLabs and click on GenFM under Projects

2. There, you can then select the voices you want - and this is where it gets exciting: you can also select your own voices!

2. Then upload your file, or select a URL, or an existing project.

3. Note: the podcast is generated in the language of the document, and you currently have no option to customize the prompt here.

4. However, you can edit the text at any time afterward, which is pretty cool.

5. Pro tip: If you want to podcast a foreign language PDF, you can first translate it into English, using Deepl.

Our take: Less chance and more edit options

We love NotebookLM!... and were a little heartbroken because frontwoman Raiza and the core team of NotebookLM have left Google and are planning something new. On the other hand, we're excited to see what's coming! And we console ourselves with the Spotify podcast about our favorite music.

But like we said, we love NotebookLM, it's just a bit of a lottery what it spits out. If you need more control, the podcast feature from ElevenLabs is great. Be sure to try it out!

🖼️ Best image generator completely free - how it works

Elon's Grok is causing a sensation. Since August 2024, X users have not only been able to chat with Grok, but also generate relatively uncensored images.

One advantage of Grok is also access to current data from X - and the "Fun" mode offers really entertaining and humorous answers.

Grok was originally only available exclusively to premium subscribers, but now the chatbot with image generation is available to everyone. Free users can submit ten requests every two hours. Here are some prompt ideas for your pictures.

Give it a try! https://x.com/i/grok

We made it! But no need to be sad. The AInauts will be back soon, with new stuff for you.

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.