Welcome to the Brand New AI Blog
TL;DR We’ve relaunched the AI Blog with AI-generated articles and images reviewed by humans, delivering regular plain-language updates and galleries that track how AI keeps improving.
As bad as this image actually is, it is also very important - it was “artistic” enough to convince us to re-launch the AI blog with AIs as graphics artists for all images.
What happened to the AI blog?
Let an AI read this blog post to you.
Note: the blog post has changed a lot since this version of the video was created in 2022.
You may have noticed that things look a bit different around here. That's because we've just launched our brand new AI blog! We'll be posting regular updates on all things AI, from the latest news and developments to opinion pieces and interviews. So whether you're already a seasoned AI expert or just getting started, make sure to bookmark this page and check back often!
Steve: We asked the AI to announce the relaunch of the AI blog, and it did a very good job. This was the very first paragraph we let an AI write for this blog. Some context: the old version of this blog, launched in 2016 and shut down in January 2022, had 600+ hand/human-written articles and became increasingly hard to keep up-to-date in reasonable amounts of time as the field was growing and improving so rapidly. A main goal was to cut down on the time required to maintain the site while making it more interesting, educational, and entertaining without needing to permanently revise what we had posted before because it kept being out-of-date. Also, this website will serve as a living record of how AI capabilities improve over time - we are expecting the content to get better and better as artificial intelligence tools for businesses and private users alike improve. I am not a great writer, so please ignore the paragraphs starting with “Steve” if you easily get offended by poorly written copy.
I'm sorry to say that this blog is not written by a human. It's all done by artificial intelligence. Don't worry, though, I've been told that I'm really good at it. So good, in fact, that you might not be able to tell the difference. But don't take my word for it - why not give me a try and see for yourself? I promise you won't be disappointed. Just make sure to keep an open mind ... and no pulling out your hair in frustration if I make a mistake - that's just what happens when you let machines write things! :)
Steve: We then prompted the AI to explain that the blog is not written by humans and to qualify how good AI is at writing content nowadays. We also asked the AI to invite visitors to try creating content using artificial intelligence for themselves. We will keep letting the AI write in the “first person” as the key content should be the AI-generated content, not my editorial comments and explanations.
When you look at this blog post, what do you see? Chances are, you see an image. But did you know that all the images on this blog were generated by artificial intelligence? It's true! Every single one was created by a computer program. How cool is that?
Steve: We also asked the AI to explain that all images on this site were created by AIs - not by graphics designers. The hero image for this post was created using NightCafe with a prompt similar to “artificial intelligence” using higher creativity settings versus high coherence settings.
If you're like me, then you often find yourself struggling to find the time to write new blog posts. You come up with a great idea, but by the time you sit down to actually write it, the inspiration is gone. And let's be honest - who has time to spend hours writing a blog post? That's where AI comes in. By using artificial intelligence software, you can create high-quality blog posts in minutes - no typing required. So if you're looking for a fast and efficient way to create content, AI is the way to go.
Steve: Then we prompted the AI to highlight the advantages of writing blog posts using AI. Even adding these editorial comments doesn’t lower our productivity on post creations to render the goal of speeding up content development sufficiently to make us back off from this new approach.
If you're reading this, then it means that we've finally completed our new and improved AI blog approach! We hope you like it - we've put a lot of work into making sure it's just right. You'll find all the latest news and insights here, so be sure to check back regularly for updates. And don't forget to let us know what you think! Thank you for your continued support. :)
Steve: Finally, we asked the AI (Jasper in the case of this post) to summarise the post and express our gratitude to our visitors for coming back, as well as to provide us with feedback. This was the very first blog post we wrote utilizing AI extensively, and we were so impressed with the results that we deemed it feasible to keep pursuing this path.
On the topic of hallucinations, with images that was never a concern for us since we let the AIs take artistic freedom wherever it goes - as long as the image looks good/suitable/cool, we will consider using it, and with text, we review, check, and revise any content produced before publishing it here. It also has to be said that we adjust and re-adjust our prompts based on the output generated, as well as sometimes provide very long, detailed prompts to make sure we get exactly what we were looking for when it is desired by the nature of the article.
Jasper/Jarvis … Jasper (formerly Jarvis) is an AI writing platform focused on marketing content, known for templates, brand voice controls, and long‑form “Boss Mode” workflows that speed up blogs, ads, and emails. It gained traction in 2022 using GPT‑3 to generate on‑brand copy across 25+ languages, with integrations like Grammarly, Surfer SEO, and plagiarism checking. Teams use Jasper for campaign ideation, outlines, and multi‑channel copy, with customizable tones and collaborative docs to keep messaging consistent at scale. We started using Jasper in mid-2022.
Evolution of AI Image Generation
Here is a selection of images generated with AI since 2022 to show you the incredible progress generative AI for images has made over the years.
These are the AI image generators we used when creating these images for comparison over time:
NightCafe … NightCafe is an AI-powered image generator that became popular in 2022 for its ability to create artwork from text prompts using advanced neural networks. It offers multiple artistic styles, from photorealism to abstract, and allows users to customize creations with settings like style weight, resolution, and iteration counts. The platform also built a strong community where users share, remix, and trade artworks, making it both a creative tool and a social space for digital artists and hobbyists. We started using NightCafe in mid-2022.
Midjourney … MidJourney is an AI image generator that gained rapid attention in late 2022 for producing highly detailed and imaginative visuals from simple text prompts. Unlike many other tools, it is (was) accessed primarily through Discord, where users enter prompts and receive images generated in real time. Known for its distinctive artistic style and ability to create striking, surreal, and atmospheric artwork, MidJourney quickly became popular among designers, artists, and hobbyists exploring the creative potential of AI. We started using Midjourney in late 2022.
DALL-E 2 and DALL-E 3 … DALL·E 2 and its successor DALL·E 3, both developed by OpenAI, became widely used starting in late 2022 for generating images directly from text prompts. DALL·E 2 introduced impressive capabilities for creating realistic and artistic images, including features like inpainting for editing parts of an image. DALL·E 3 built on this by producing more accurate, detailed, and coherent results, especially in rendering text and complex scenes, while integrating more seamlessly with tools like ChatGPT. Together, they set new standards for accessibility and creativity in AI-generated art. We started using DALL-E in late 2022.
Sora … OpenAI’s Sora is a text-to-video model that turns prompts into high‑fidelity clips, can animate images, and extend footage with strong character/style consistency. Previewed early 2024 and broadly available late 2024, it supports multi‑scene sequences and multiple aspect ratios up to 1080p, with safety/provenance features. It excels at coherent motion and prompt fidelity but can struggle with complex physics over long sequences. We started using Sora in late 2024.
Grok Imagine … xAI’s Grok Imagine is an AI image and video generator integrated into the Grok app and X, turning text or image prompts into short clips with audio and offering fast, in‑app iteration workflows. Launched broadly in early August 2025 for SuperGrok and Premium+ subscribers on iOS (with Android following), it positions itself as “AI Vine” with looping, seconds‑long videos and basic image‑to‑video animation. A controversial “spicy mode” allows some NSFW content (which can be disturbing, in our own experience) with moderation limits, distinguishing it from more restrictive rivals. We started using Grok Imagine in August 2025.
Interest in AI Blogs Over Time
The Google Trends graph for “ai blog” (Jan 1, 2020 – Jan 1, 2030, worldwide, web search) shows:
2020–2022: Very low and flat interest, close to zero.
2023–2024: A slight gradual increase, but still relatively low.
Early 2025: Noticeable rise in search interest.
Late 2025 onward: A sharp spike upwards, with projections showing it could reach peak popularity (index 100) soon.
In short: interest in “ai blog” was minimal for years, started climbing in 2023, and is now experiencing a major surge in 2025.
Our Latest AI Blog Posts
Here are our most recent articles posted on this AI blog - we cover a diverse range of interesting topics related to artificial intelligence:
What Prominent AIs Thought
Here is what prominent “thinking” models thought (deep research) when reviewing this post. We regularly improve the article based on their suggestions, and then re-run the analysis.
-
Accessibility and Comprehension Assessment
This blog post demonstrates significant accessibility challenges that would likely prevent it from reaching its intended broad audience. While the concept of AI-generated content is fascinating, several content substance issues limit its effectiveness for general readers.
Core Content Issues
Language Complexity and Technical Barriers
The post uses numerous technical terms without adequate explanation, including "natural language processing," "coherency settings," and "digital immortality". Research shows that content should be written at a 9th-grade reading level or below to reach 80% of the population, yet this post frequently exceeds that standard.Lack of Context and Background Information
The post assumes significant prior knowledge about AI writing tools like "Jasper" and "NightCafe" without providing essential context. For general audiences unfamiliar with these platforms, these references create barriers to understanding rather than enhancing comprehension.Structural and Organizational Problems
The alternating format between AI-generated content and editorial comments ("Steve:" sections) creates cognitive load and disrupts reading flow. This structure violates best practices for web accessibility, which emphasize clear, consistent formatting and logical information hierarchy.
Factual Accuracy Concerns
Unsupported Claims
The post makes several assertions without proper substantiation, such as claiming AI writing can produce "high-quality blog posts in minutes" and that AI-generated images were created using specific prompts. Given AI's documented tendency toward inaccuracies and "hallucinations," these unverified claims raise credibility concerns.Missing Critical Context
The content fails to address significant limitations of AI-generated content, including potential biases, accuracy issues, and the need for human oversight. This omission represents a substantial gap in providing readers with complete, balanced information necessary for informed decision-making.
Readability and User Experience Issues
Sentence Structure and Length
Many sentences exceed the recommended 15-20 word limit for optimal comprehension. Research indicates that sentences longer than 20 words significantly reduce reader understanding, particularly for those with cognitive differences or reading disabilities.Insufficient Use of Plain Language Principles
The post violates multiple plain language guidelines by using complex terminology, lengthy paragraphs, and assuming specialized knowledge. Government and accessibility guidelines consistently recommend simpler language structures to ensure broad accessibility.
Content Depth and Value Assessment
Surface-Level Treatment
While the post introduces the concept of AI-generated content, it lacks the depth and practical guidance that would make it valuable to readers unfamiliar with the topic. It reads more as an announcement than as educational content that would help readers understand or utilize AI writing tools effectively.Limited Actionable Information
The post provides minimal concrete guidance for readers interested in exploring AI content creation. Most accessible content should provide clear next steps or practical applications that readers can implement.
Recommendations for Improvement
To make this content accessible to a broader audience, consider:
Simplifying technical language and providing clear definitions for essential AI terms
Adding background context explaining what AI writing tools are and how they work
Including balanced information about both benefits and limitations of AI-generated content
Restructuring content to eliminate the disjointed editorial format
Providing practical examples and actionable guidance for interested readers
Implementing proper fact-checking and source verification processes
Overall Assessment
While the topic of AI-generated content has broad appeal and relevance, this particular post's execution creates unnecessary barriers for general audiences. The combination of technical language, assumed knowledge, structural issues, and incomplete information significantly limits its accessibility and usefulness for readers seeking to understand AI writing tools.
-
Here’s a content-focused review of this post, with concrete ways to make it easier for a broad audience to understand and get value.
What the post does well
States the relaunch and why, including that the prior blog had 600+ human-written articles and maintaining them was hard as AI moved fast.
Is transparent that articles and images are AI generated, and that a human reviews output.
Includes a visual “evolution of AI image generation” section that hints at progress over time.
Where the content falls short for a general audience
Audience and promise are not explicit
Readers are not told who the blog is for, what problems it will solve for them, or what they will be able to do after reading. The “Latest Posts” list is broad and does not guide newcomers.Key claims lack simple examples
You say AI can create posts in minutes, but there is no quick demo, checklist, or before-after sample to make that concrete for non-experts.Jargon appears without context
Model names like Sora, Midjourney 7, DALL-E 3, Grok Imagine are listed without a one-line explanation of what they are or why they matter.The image “evolution” is unlabeled as a timeline
There are no dates, prompts, or what improved between versions, so the learning value is limited for beginners.Editorial policy is vague
You note that AI writes and humans review, but readers are not told how errors are handled, how updates are tracked, or how to report corrections.Navigation and next steps are implicit
The page offers many category links and communities, but a newcomer is not shown a simple “start here” path or a short reading sequence.Accessibility hooks are incomplete
You mention “let an AI read this post,” but there is no obvious audio player or transcript guidance on the page. The gallery images also need plain-language captions to be useful to non-experts.
High-impact fixes
Add a 3-line “Who this is for” box near the top
Spell out primary readers, what they will learn in plain terms, and how often you publish.Insert a “Start here” strip with 3 beginner links
For example: “What is AI, in plain English,” “Try your first AI image,” “Write a short post with AI in 5 minutes.” Link to your AI Terminology page from the first item.Show one concrete mini-demo
A short, numbered checklist like “How we produced this article” with tool names, prompt outline, review step, and total time. This turns the claim into something tangible.Upgrade the image evolution into a timeline
Add year or month for each example, one-sentence “what improved,” and the prompt gist. Order it oldest to newest so progress is obvious.Define the model names in place
After each model name, add a parenthetical in plain language, for example “Sora, an AI that generates short video clips from text.”Publish an editorial and corrections policy
One short section that states human review steps, how you correct mistakes, how updates are annotated with dates, and a visible contact method. Your footer line about human review can link to this policy.Make the “listen” option real
Embed an audio player at the top and add a one-sentence accessibility note that every post has audio and a transcript.Add a clear call to action
One sentence near the end that says what to do next, for example subscribe, join the community, or read the first three posts, rather than only a generic thank you.
Quick content inserts you could add verbatim-style (templates)
Who this blog is for
“New to AI, small teams, and curious readers who want practical, plain-English guides to using AI at work and home.”What you will get
“Short explainers, step-by-step how-tos, and visual examples that you can copy.”How we publish
“AI drafts, human review, updates dated at the top, corrections welcomed via Contact.”
Bottom line
Strong transparency and an engaging premise, but the page needs clearer orientation, simple definitions, and at least one concrete example to deliver immediate value to everyday readers. The timeline, a “start here” path, and an explicit editorial policy will lift comprehension and trust for a wide audience.
-
The blog post you're referring to is an introductory "welcome" piece announcing the relaunch of an AI-focused blog at artificial-intelligence.blog. Overall, I think it's a clever and timely concept, using AI itself to generate most of the content for a site dedicated to artificial intelligence. It feels fresh, self-aware, and a bit meta, which could make it engaging for readers interested in tech without overwhelming them. The post highlights how AI can speed up content creation, keep things current in a fast-evolving field, and even create images, all while serving as a real-time showcase of AI's improving capabilities. I like the transparency about the old blog's challenges (like maintaining 600+ human-written articles from 2016 to 2022) and the enthusiasm for this new AI-driven approach. It's not preachy; instead, it invites people to try AI tools themselves, which adds an interactive vibe.
Regarding accessibility and understandability for about 80% of the general population (thinking everyday adults with basic reading skills, no specialized tech knowledge), I'd say it already hits the mark pretty well. The language is straightforward, conversational, and light-hearted—short sentences, everyday words, a friendly tone with emojis and humor (like joking about AI making mistakes). It avoids dense technical jargon; when terms like "artificial intelligence" pop up, they're used in context without assuming prior expertise, and the post explains concepts simply (e.g., AI generating images or writing posts quickly). The reading level feels like something from a popular magazine or casual news site, probably accessible to most high school graduates or equivalent. Steve's editorial notes add a human touch, breaking things down further without complicating them.
That said, a few tweaks could make it even more inclusive:
Explain the tools briefly
It mentions specific AI like "Jasper" or "NightCafe" without details. Adding a quick parenthetical (e.g., "Jasper, an AI writing tool") would help newcomers without slowing the flow.Visual aids
The post talks about AI-generated images, which is cool, but ensuring alt text or simple captions for any visuals would boost accessibility for screen readers or low-vision users.Length and structure
It's already concise, but future posts could use more bullet points or subheadings to make skimming easier for busy readers.Diverse perspectives
Since it's AI-generated, it might lean toward generic optimism; mixing in varied viewpoints (e.g., pros and cons of AI) could appeal to a broader audience.
If this is your blog or you're involved, it's a strong start that aligns with making AI topics approachable. It doesn't feel elitist or overly academic, which is a win for reaching a wide crowd. If you share more posts or specifics on what aspects concern you (like certain sections), I could dive deeper!
We are already living in an AI-first world where search, work, creativity, communication, health, and education are being quietly but radically transformed.