<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Below Average Writer - Medium]]></title>
        <description><![CDATA[Between spell check and an AI-written article is the below average writer with great ideas. - Medium]]></description>
        <link>https://medium.com/below-average-writer?source=rss----dfd14cd7d946---4</link>
        <image>
            <url>https://cdn-images-1.medium.com/proxy/1*TGH72Nnw24QL3iV9IOm4VA.png</url>
            <title>Below Average Writer - Medium</title>
            <link>https://medium.com/below-average-writer?source=rss----dfd14cd7d946---4</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 21:08:13 GMT</lastBuildDate>
        <atom:link href="https://medium.com/feed/below-average-writer" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[AI Slop Meets Skeptic]]></title>
            <link>https://medium.com/below-average-writer/ai-slop-meets-skeptic-2001a5c4998d?source=rss----dfd14cd7d946---4</link>
            <guid isPermaLink="false">https://medium.com/p/2001a5c4998d</guid>
            <category><![CDATA[notebooklm]]></category>
            <category><![CDATA[editing]]></category>
            <category><![CDATA[gemini-notebook]]></category>
            <category><![CDATA[ai-slop]]></category>
            <category><![CDATA[innovation4est]]></category>
            <dc:creator><![CDATA[Ray Daly]]></dc:creator>
            <pubDate>Wed, 09 Sep 2026 15:12:08 GMT</pubDate>
            <atom:updated>2026-09-09T15:12:08.862Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*W8FhYM1u9SZGQeiZWhpf2Q.jpeg" /></figure><h4>Other parts of writing are helped by AI</h4><p>As a below average writer, I’ve spent hours reading and listening to arguments about AI Slop. It is a nuanced topic.</p><p>First, I think AI Slop is probably a short term problem. In ten years people will not understand it in the same way as we do now. And obviously, ten years prior we did not have a concept of it. This is a cultural component.</p><p>Second, it maybe that AI Slop is a phase of AI that is obsolesced already. As I listen to AI <a href="https://www.youtube.com/@NateBJones">podcasts like Nate B Jones</a>, they have moved on to the agentic age of AI. Their concern is the <a href="https://www.youtube.com/watch?v=xOi5nDH0lu0">Hugging Face break-in</a> and not slop. Obsolesced is not obsolete but a phase in the lifecycle of an innovation (see Laws of Media). Soon after magazines entered the obsolesced phase there were more magazines published than ever before. AI Slop writing may have another life.</p><p>Third, deteccting AI Slop has become an industry and a trend that is very alarming to me. I too tested <a href="https://www.pangram.com/">Pangram</a>, an AI detector, after listening to <a href="https://www.youtube.com/watch?v=HdU7OAsdViA">Charlie Warzel’s podcast on “Did a Human Write This?”</a><em>. </em>My opinion on AI detectors<em> </em>probably requires an entire article after doing more research.</p><p><a href="https://www.youtube.com/watch?v=1dYp9ymqy_g">Steven Johnson podcast with Dan Blumberg</a> on writing and cognitive uploading was the catalyst for me writing and designing the system for <a href="https://medium.com/below-average-writer/using-gemini-notebook-from-drafting-to-writing-c31ad9309c23">Using Gemini Notebook from Drafting to Writing</a>. That work is ongoing since creating an article is more than a draft becomes an article.</p><p>As many of the AI Slop podcasts have pointed out that professional writers have editors. These wrtiters love a good editor. My digital system does not have an editor so at the end of the process a human editor goes to work. There are writing tools to help with that part of writing. More are in development.</p><p>So this is a long introduction to <strong>thank Mihailo Zoin</strong> for his article: <a href="https://medium.com/@kombib/gemini-notebook-analyze-articles-six-prompts-7f7bbadd9fa6">Six Gemini Notebook Prompts That Read an Article Like a Skeptic</a>. His six prompts are for use within Gemini Notebook. He calls the action “a skeptic” but I used it like a type of editor for my recent article on <a href="https://medium.com/innovation4est/the-innovation-inventory-experiment-fec4b65371bd">The Innovation Inventory Experiment</a>. I rolled up the prompts into one prompt and used it in some of the last rewrites. It gave me a much stronger article than what I started four or five years ago. In paticular, it lead me to pull an entire section out of that article for clarity and it will become a new article.</p><p>I think AI has a place in writing between the spell checker and writing the full article. The “skeptic” prompts worked for me.</p><ul><li><a href="https://medium.com/below-average-writer/using-gemini-notebook-from-drafting-to-writing-c31ad9309c23">Using Gemini Notebook from Drafting to Writing</a></li><li><a href="https://medium.com/@kombib/gemini-notebook-analyze-articles-six-prompts-7f7bbadd9fa6">Six Gemini Notebook Prompts That Read an Article Like a Skeptic</a></li><li><a href="https://medium.com/innovation4est/the-innovation-inventory-experiment-fec4b65371bd">The Innovation Inventory Experiment</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2001a5c4998d" width="1" height="1" alt=""><hr><p><a href="https://medium.com/below-average-writer/ai-slop-meets-skeptic-2001a5c4998d">AI Slop Meets Skeptic</a> was originally published in <a href="https://medium.com/below-average-writer">Below Average Writer</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Using Gemini Notebook from Drafting to Writing]]></title>
            <link>https://medium.com/below-average-writer/using-gemini-notebook-from-drafting-to-writing-c31ad9309c23?source=rss----dfd14cd7d946---4</link>
            <guid isPermaLink="false">https://medium.com/p/c31ad9309c23</guid>
            <category><![CDATA[gemini-notebook]]></category>
            <category><![CDATA[notebooklm]]></category>
            <category><![CDATA[writing-tools]]></category>
            <category><![CDATA[ai-writing]]></category>
            <category><![CDATA[gemini]]></category>
            <dc:creator><![CDATA[Ray Daly]]></dc:creator>
            <pubDate>Wed, 19 Aug 2026 19:26:29 GMT</pubDate>
            <atom:updated>2026-08-19T19:26:27.695Z</atom:updated>
            <content:encoded><![CDATA[<h4>Creating your story</h4><h4>The “Ongoing Source” Strategy to Keep Your Voice</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ttRJm_lC_J_Z-t76U1D3gw.jpeg" /><figcaption>Three tab layout for Ongoing Source: Gemini Notebook, Ongoing Chats, and Working Document generated by Gemini</figcaption></figure><p>When using Gemini Notebook it is always tempting to generate a few more overviews, create a couple more slideshows, and in particular type in a few more prompts in the chat box.</p><p>Before you know it, the generated chat text is longer than the original source. As you interact, the conversation naturally fills up with a mix of “chaff” like conversational fluff, half baked ideas, and abandoned trial balloons mixed in with actual “gold nuggets” of insight.</p><p>You may feel defeated. So you think I’ll get the AI to summarize that long chat which just makes the wall of text longer. But you don’t want to delete the entire chat because you had a good exchange and using just the summary may miss some things.</p><p>The problem is that raw AI chats are append-only and cannot be edited. When faced with this lengthy wall of text, relying solely on standard keyword search (Cmd+F) to retrieve your insights quickly doesn’t work. The whole thing becomes a messy bottleneck. For me, this is the original AI slop.</p><p>Nate B. Jones laments about the problem in his podcast “<a href="https://youtu.be/AWGoOtNgw3c?si=-skxUuZdh4x7P9v-">AI News &amp; Strategy Daily</a>: ”</p><blockquote><em>Sure, before we had the blank page, but now we have what’s arguably worse: acres and acres and acres of pages that maybe nobody has read that someone’s expecting us to pay attention to.”</em></blockquote><p>So I needed a solution tailored to Gemini Notebook and which took advantage of Notebook’s features like the closed environment of grounded sources.</p><h3>Three tab solution</h3><p>My solution when using a Notebook for drafting and writing is to add a new Google doc called “Ongoing Source Chats.” In my browser I keep the Notebook open in one tab, the document i’m writing in another, and the ”Ongoing Source Chats” document in a third tab. When I have chats I want to save, I copy them into the Ongoing Chats doc. After reviewing the entire chat, I delete the history in the chat panel. The result is a blank page for chatting. This is the three tab solution.</p><p><strong>But wait!!!</strong> If the idea is to preserve chats, why delete it? It is not really deleted because the edited chat history is now in your Ongoing Chats document which is added as a source for your notebook. You preserved the best of your chats and deleted the chaff.</p><h3><strong>Making it work with “Slash-and-Burn”</strong></h3><p>Make certain you avoid saving all the chats in the Ongoing Source document. Skip the conversational fillers. Select only those responses that provide good insights, decisions, or useful snippets.</p><p>For example, imagine you prompt the AI to provide three possible solutions to a problem. Or three possible labels. You rarely need all three permanently on the record. Instead of copying the entire verbose output, you simply copy the single winning response and paste it into your Ongoing Source doc.</p><p>In fact, this exact article is a perfect example. Early in the writing process, we brainstormed three different options for how to save AI chat outputs (saving to a Studio note, saving to a new doc, or using a canvas). Rather than dumping that entire raw, exploratory chat, we used the “slash-and-burn” method to extract only the final “Ongoing Source” strategy and its core pros and cons.</p><p>Once you have captured the best of the exchange like this, you can safely delete the raw chat history from the Notebook without losing your progress.</p><h3>Formatting for Semantic AI Retrieval</h3><p>The Ongoing Chats document can grow to thousands of words and hundreds of responses. I find it useful to have a timestamp embedded into each response, especially in a format that is AI friendly.</p><p>As you can see my timestamps are formatted; Ref: [UNIX_TIMESTAMP] | YYYY-MM-DD HH:MM:SS EDT. The UNIX timestamp makes it easier for Gemini to retrieve responses. As an example you might want to look at what you did last Tuesday.</p><p>To add timestamp click the <strong>Configure Chat</strong> icon (the tuning sliders next to the prompt box) in Gemini Notebook. Then set the goal to <strong>Custom</strong>. Finally add a standing instruction like:</p><blockquote><em>CRITICAL FORMATTING RULE: Every response MUST begin on the very first line with a timestamp: ‘Ref: [UNIX_TIMESTAMP] | YYYY-MM-DD HH:MM:SS EDT’</em></blockquote><p>This ensures every response, saved note, and copied session block has precise temporal grounding before it syncs back into your primary sources. And I like being able to extract a development timeline.</p><h3>The Auto-Sync Advantage</h3><p>A key to making Ongoing Sources work is the relatively new feature where Google Doc files are automatically synced. Make a change to the doc and it is periodically synced with Notebook in the background. If needed you can force an immediate sync with a panel refresh.</p><p>The true magic of chats as sources lies in how Gemini Notebook handles Google Docs. As you paste your refined chat notes into the document, Gemini Notebook continuously and automatically indexes it. There is no need for manual file re-uploads or managing dozens of cluttered files. Your newly pasted insights become ground-truth context for your future queries.</p><h3>Selecting Sources Advantage</h3><p>Let’s use the Ongoing Chats Source now that your old chats were deleted. The power user move is where you select and deselect sources in the source panel. This allows you to have chats with just your edited chat history (State A), or with just your research sources (State B), or both (State C).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LrTAPSfa3vb2cJ7Q9GKKcg.jpeg" /></figure><h3>State A: Isolated Reflection Mode</h3><p>Only “Ongoing Chats” Selected. By excluding external reference sources, Gemini focuses on past sessions. Below are examples of prompts you might consider in this state:</p><ul><li>“Based on my ongoing notes from this month, what were the main challenges with…”</li><li>“How has my approach evolved over the past month?”</li><li>“List every trial balloon marked #unresolved.”</li><li>“Generate a study guide (or FAQs or Briefings) built <em>entirely</em> out of your raw thought history.</li><li>“Identify the three most frequent #ideas I’ve documented over the past two weeks.”</li><li>“What did I forget to consider?”</li><li>“What issues are unresolved from the past month?”</li></ul><h3>State B: Clean Ground Truth Mode</h3><p>Uncheck Only “Ongoing”. By excluding the “Ongoing” lets you work just with your writing and researched sources. This is a good state to verify the facts against the sources and look for missing sources.</p><ul><li>“Do we have a source for all of the facts presented?”</li><li>“Draft a list of references.”</li><li>“Verify the following claim against the primary research data to ensure there is no AI-generated noise.”</li><li>“Would any source disagree with our conclusion?”</li><li>“Extract all the statistical data points mentioned in the report.”</li><li>“Is one source doing too much work?”</li><li>“Can you suggest a quote from the source about ….?</li><li>“What did this author have to say about this case?”</li></ul><h4><strong>The Power of the “Inverted Search”</strong></h4><p>In State B, AI provides a powerful option that Steven Johnson calls the “Inverted Search”. Instead of searching for keywords that <em>are</em> there, you ask the AI to scan your selected sources to find what you <em>left out</em>:</p><p><em>“I have just written this draft. Scan all of my selected reference documents and tell me what I may have left out.”</em></p><h3>State C: Integrated Mode</h3><p>Selecting all sources including your past chats is great for a writing session especially if past sessions need to inform the current session.</p><ul><li>“Was that my idea or did AI suggest that?”</li><li>“How do the ideas I brainstormed in my ongoing notes compare to the official guidelines in the primary source document?”</li><li>“Using the theoretical framework from the textbook, evaluate the trial balloons we discussed last week.”</li><li>“Where does my ongoing project outline conflict with the facts presented in the reference PDFs?”</li><li>“Synthesize a summary combining the author’s main argument with my reflections from Tuesday’s session.”</li></ul><p>Using the above modes you assure that the document is yours. As you use these modes to interact with ai, <strong>you</strong> can directly edit your working document in the other tab or window. And obviously you can ask ai to review your revised document and iterate more. But it is your edits, your review, and your document.</p><p>While you could have the AI directly modify the working document, but then your voice trails off. Keep the review in the chat and then in the document tab revise your document. By keeping the review in the chat also creates the historical record.</p><p>As Steven Johnson said in the podcast with Dan Blumberg: <a href="https://youtu.be/1dYp9ymqy_g?si=0bKDtxEkHHY81rZF">How Google NotebookLM’s Steven Johnson uses AI as a second brain</a></p><p><em>“And it’s a sense of </em><strong><em>you have to know what you want to say generally</em></strong><em>, and that takes separate thinking and outlining and wrestling with the ideas… Is it inappropriate to ask the AI to help you step by step to generate the first drafts if you are guiding it at every stretch of the process?”</em></p><h3>Continue Drafting to a Final Edit</h3><p>At this point you continue to edit and chat. There are two different options to save your chats. The first is to immediately cut and paste useful responses into the Ongoing Chat document. The second is to use the Notebook button to “Save to Note.” Then at the end of the session you can select which “Chat Notes” to cut and paste into Ongoing chats.</p><p>I find myself using both options. ”Save to a Note” is particularly useful if you want to stay in the writing “flow” and don’t want to be distracted. I also have found it useful when testing out an idea. In either case, at the end of the writing session I’ll review all the “Chat Notes” and selectively “Slash and Burn” or “Cut and Paste” the notes into the Ongoing Chat document.</p><h3>The Studio Advantage</h3><p>After many iterations your drafting is nearly complete. At this point I take advantage of Gemini Notebook’s Studio panel. First, I’ll select the working document as the source. Then I get Notebook to generate an Audio and Video Overview, usually with a focus of “Is the source appropriate for my audience.” Or maybe I’ll do a Mind Map to make certain I’ve covered the topic I’ve wanted. And finally I’ll generate the Audio Overview critique as one final pass. Or maybe even a second final pass.</p><p>These artifacts are similar to having another “set of eyes” to review your document. In particular when I hear the overviews and back away from the typing, I get some perspective on what I’ve rendered.</p><h3>Conclusion — The Polished Document</h3><p>The Ongoing Source strategy is useful for writers who want to use AI yet do the work to make the ideas and the writing theirs.</p><p>After all the iterations, the chat reviews, and studio artifacts, you’ll be proud of <strong>your</strong> work. While appreciating AI’s help in polishing the document, at the end it is your writing making your points in your voice.</p><h3>Afterthought: Figure Grounded</h3><p>Our Innovation Forest articles conclude with a feature called Figure Grounded. The prompts are provided for further exploration of a topic.</p><p>In the Notebook chat panel try asking <em>“The ongoing chats have a timestamp for each response. Suggest ten ways to take advantage of that with corresponding prompts.”</em></p><p>I’d be interested if you found any of the suggestions useful in your writing. Please leave a comment about your experience.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c31ad9309c23" width="1" height="1" alt=""><hr><p><a href="https://medium.com/below-average-writer/using-gemini-notebook-from-drafting-to-writing-c31ad9309c23">Using Gemini Notebook from Drafting to Writing</a> was originally published in <a href="https://medium.com/below-average-writer">Below Average Writer</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>