I refer to the «Cambrian explosion of artificial intelligence» as the rapid proliferation of large language models like GPT and the explosive growth of AI services based on their advances.
AI tools to try out
This list is by no means comprehensive and I won’t be able to keep up with the development in the field. But it may give you a good starting point.
A prompt is the text that the AI should use to generate something.
Text to image – generate images based on text prompts
- DALL-E 2 – image generator for realistic looks
- Dream Studio – text to image with Stability AI
- Flair – text to background image of product photos
- Illustroke – text to vector images. You «only» get 3 variations per prompt, hence the limited match with your needs. Two for free.
- Midjourney – prompt to image generator for artistic looking graphics
- Stable Diffusion – text to image or video
- Stockimg.ai – text to photo or logo. One for free. This is too little to test compared to the free amount you get on Midjourney.
Text to text – create new or edit existing copy
- Bertha.ai – marketing copy that drives conversions | WordPress plug-in
- Copy.ai – copy writing for marketing. In my short test I got less good results than with Deepl Write.
- Copymatic.ai – generates copy images
- Deepl Write – correct copy or rewrite it. The tool is just as convincing as its translating counterpart Deepl Translate.
- Frase.io – copy writing for search engines | WordPress plug-in
- GoCharlie – copy writing for SEO and image generation
- Ink
- Jasper.ai – copy writing and art image creation
- Kafkai – copy writing for SEO
- Neuroflash.com – copy writing in German
- Ocoya – create & schedule social media content
- Pencil – produces engagement-driven promos for ads
- Perplexity
- Rytr
- Shortly – expand on a text fragment
- Smartwriter – writes personalized emails to prospects
- Texta – search-optimized blog posts
- TLDRthis – summarize copy in a click. In my quick test I didn’t get the quality I expected.
- Tribescaler – create better ad hooks
- Wordtune – rewrite copy for max impact, lists sources
- YouChat – much like ChatGPT, YouChat provides copy and code. But it offers better availability as it is less well known. Discloses sources for its answers. Offers Chrome plugin.
Text to slides
Text to video or voice and music – compile videos or audio based on text prompt
- Beatoven – create unique royalty-free music. Their teaser video is not particularly convincing. I haven’t tried the tool just yet.
- BHuman – AI personalized videos
- Descript – clone your voice for text to speech
- Maverick – Generate personalized videos at scale
- Murf – text to voice
- Pictory.ai – text to video
- Synthesia – compile AI-based videos
- Vidyo – create short-form videos from long-form copy. Trial comes with 1h 15 minutes free processing time. Doesn’t seem to work with videos without spoken audio.
Text to 3D model
- Mirage – generate 3D model from text
Thought to x
- Neuralink – Elon Musk’s brain reading startup
How we got here
For some time now, «smart» systems such as the excellent translation service DeepL or Google Photos have been preparing us for the new possibilities that machine learning and artificial intelligence bring.
When the company Open AI presented its image generator Dall-E in the fall of 2022, the tech world was astonished: The images generated by the AI based on a short text prompt are of excellent nature!
When the next image generator Midjourney also became known, the wave of enthusiasm no longer gripped only tech journalists, but also a broader public: Midjourney’s images are even more creatively valuable, see my header image.
Open AI went on to cause a further sensation in late 2022 when it made its chatbot ChatGPT public: The chatbot was already so good in version 3.5 that millions of users worldwide wanted to experience it for themselves («GPT» stands for «generative, pre-trained transformer).
Within a few weeks (!), first the review columns of tech publications, later also of popular media, were filled with reports about the chatbot’s comprehensive knowledge in all possible fields.
In my very first test, I had ChatGPT program a grid of cubes in two different 3D modeling tools. I had to laugh out loud at the result: it was staggeringly good! See the results in my review.
The race for chatbot supremacy
Since the beginning of 2023, a whole wave of new AI applications has been emerging on a daily basis.
Microsoft’s announcement to integrate the functionality of ChatGPT into their products and their search engine Bing has launched the race for chatbot supremacy: Apparently we won’t be consulting search engines for much longer, but will be getting advice from chatbots.
Much like Hiro Protagonist in Neil Stephenson’s metaverse sci-fi novel «Snow Crash».
If you haven’t already, I recommend the reading: Neal teaches many interesting topics as well as all things Metaverse. I read the book twice for its Metaverse aspect (Read more about my own forays into the Metaverse).
Google is responding by integrating DeepMind’s AI into their own chatbot, «Sparrow.» They also have the allegedly sentient chatbot «LaMDA» in the pipeline.
I call it the Cambrian Explosion of AI in 2023
With the «Cambrian Explosion» of Artificial Intelligence I refer to the rapid advancement and proliferation of large language models like GPT and the countless services built on top of them.
These models have the ability to generate a wide range of outputs, from creative writing to code, based on a simple text prompt.
One of the main reasons for this explosion is the availability of vast amounts of data and computing power. With more data to train on, models can become more accurate and versatile in their outputs.
Additionally, the development of powerful hardware, such as graphics processing units (GPUs), has made it possible to train these models in a relatively short amount of time.
Another important factor is the open-source nature of many of these models. This means that anyone can access the code and train their own version of the model.
This has led to a proliferation of different versions of GPT, each with its own strengths and weaknesses.
The impact of this Cambrian Explosion has been significant: Artificial intelligence services are now able to generate a wide range of outputs, from creative writing to code, that were previously unimaginable.
As of 2023, this has led to a number of exciting new applications, such as the use of GPT to generate poetry, music, art, and even entire comics and books.
In addition to these creative applications, GPT has also been used to generate code, advice, and instructions. This has the potential to greatly improve efficiency in industries such as software development and customer service.
However, it’s also important to note that with the increasing capabilities and availability of these models, it’s crucial to also address concerns around privacy, security, and ethical use.
Even taxation of AI systems may eventually become a topic in order to finance a universal basic income for knowledge workers who lost their jobs to AI automation.
The AI Revolution was predicted in 2015
In 2015, Tim Urban of Wait But Why published a blog post in which he predicted the rapid advancement of artificial intelligence, which he referred to as the «AI Revolution.»
Urban argued that developments in AI would accelerate rapidly, leading to a future in which computers would surpass human intelligence.
As of the year 2023, we are now entering the predicted «hyper quick deployment» phase of AI: With ChatGPT leading the way at the end of 2022, in 2023 a plethora of services emerged around large language models.
In summary, the Cambrian Explosion of Artificial Intelligence, represented by the proliferation of large language models like GPT, has led to a wide range of exciting new applications and possibilities.
However, it’s important to also address concerns around the ethical use and implication of these models.
Mankind was not prepared for this, at least not the knowledge workers.
2 Kommentare
Hi, Walter, I guess I won’t be translating any more of your engaging travel blogs. You and your new AI pal are pretty good, as evidenced by this blog post. I’ll be applying for that basic income for displaced translators!
Best wishes for the brave new world of AI.
Your human pal,
Henry
Dear Henry
Yes, service providers no longer have to offer «conventional» translations.
But you offer very special custom translations, which a machine can’t imitate so easily ;-)