Techstravaganza

Productivity and Creativity with AI, Wednesday, Sept 2, 2026

Written by

Macara

in

Hello Techstravaganza friends,

Siena and Ali provided a wonderful demonstration and discussion of the ways to use AI for productive and creative tasks. Have a listen at the following link: https://open.spotify.com/episode/4w5UC2h0um5oa0Mq5FaXFJ

Show Notes

Workshop Introduction and Session Setup

Siena and Ali opened the session, welcomed new and returning members, and outlined the plan to compare three AI assistants — ChatGPT, Google Gemini, and Perplexity — on productivity and creativity tasks. 

  • Details
    • Siena: Introduced herself as host and co-facilitator alongside Ali, welcomed all attendees, and explained the session focus on AI for productivity and creativity.
    • Ali: Noted plans to open the floor for questions around 2:30 PM and flagged topics from the Textravaganza Workshops WhatsApp/iMessage group, including questions about computers and iOS betas.
    • Siena: Confirmed an alarm was set for 2:30 PM and sent a message to the Techstravaganza Workshops group to inform members of the session agenda.
  • Conclusion
    • The session was structured as a live demonstration followed by open Q&A.
    • Notes were being captured via recording, managed by Lily.

AI Tool Comparison — PDF Summarization

Siena demonstrated uploading an inaccessible, image-based PDF (a MIDI Plus Mini Engine USB user manual) to each AI assistant and requesting a summary. The PDF was scanned and not readable by screen readers directly.

  • Details
    • Siena: Explained that image-based PDFs cannot be read by screen readers such as JAWS, and that she currently uses DocuScan Plus for this task. She wanted to test whether AI could replace or supplement that workflow.
    • Siena (on Gemini): Uploaded the PDF via Google Drive clipboard paste. Gemini produced a well-structured summary with headings (Product Overview, Key Features, Power and Battery, Controls and Interface), making it easy to navigate by heading using JAWS.
    • Siena (on ChatGPT): Uploaded the same file. ChatGPT also produced a concise summary with headings and bullet points. Siena noted Gemini’s summary felt slightly more concise.
    • Siena (on Perplexity): Uploaded the file and requested a summary. Perplexity produced a usable summary with headings, though it prompted upgrade notices during the process.
  • Conclusion
    • All three AI assistants successfully summarized the PDF manual.
    • Gemini’s output was considered slightly more concise and well-structured.
    • Perplexity’s free plan showed limitations with repeated upgrade prompts.

AI Tool Comparison — Verbatim Text Extraction from PDF

Siena asked each AI to extract the full text of the manual exactly as written, without summarizing, to test whether the tools could serve as an accessibility workaround for inaccessible documents.

  • Details
    • Siena (on Gemini): Required multiple, increasingly specific prompts before Gemini extracted the English text correctly. Initially, Gemini attempted to summarize. After specifying “extract only the English text exactly as written,” Gemini eventually succeeded but took significantly longer.
    • Siena (on ChatGPT): With a single clear prompt, ChatGPT extracted the full verbatim English text quickly and accurately, preserving the original wording including grammatical errors present in the source document.
    • Siena (on Perplexity): Perplexity extracted some text but the output was disorganized, with repeated website URLs, jumbled characters, and text that was difficult to read with a screen reader.
    • Liz: Asked for clarification on what “productivity” means in this context. Siena explained it refers to tasks that increase efficiency, such as making inaccessible documents readable.
    • Liz: Shared a personal use case — using Meta AI glasses to have jokes and memes explained to her, which she found very helpful.
    • Siena: Confirmed this is a valid and practical use of AI, noting she had used ChatGPT and Gemini to interpret memes sent by family members.
  • Conclusion
    • ChatGPT performed best for verbatim text extraction with minimal prompting.
    • Gemini required more specific prompting but ultimately succeeded.
    • Perplexity’s output was difficult to read and considered a partial failure for this task.
    • Prompt specificity is critical, especially with Gemini; users may need multiple attempts to get the desired output.

AI Tool Comparison — RTF File Conversion

Siena asked all three AI assistants to convert the uploaded PDF into an RTF document that could be opened and read in a word processor using a screen reader.

  • Details
    • Siena (on ChatGPT): ChatGPT converted the file to RTF within seconds and provided a download link. The resulting file opened in LibreOffice and was readable, though it contained some RTF formatting artifacts (backslashes and escape characters).
    • Siena (on Gemini): Gemini packaged the RTF file inside a ZIP archive for “secure downloading.” After extraction, the file opened cleanly in LibreOffice with readable, well-formatted text. Siena declared Gemini the winner for this task.
    • Siena (on Perplexity): Perplexity prompted an upgrade to Pro for document creation but still produced a downloadable RTF. The resulting file contained extraneous words and was difficult to read, earning a “half-pass.”
  • Conclusion
    • Gemini produced the cleanest RTF output and was judged the best performer for this task.
    • ChatGPT’s output was functional but contained formatting artifacts.
    • Perplexity’s output was the weakest and required a paid plan for full functionality.

AI Tool Comparison — Creative Song Generation

Siena shifted to a creativity task, asking each AI to generate a country song about the frustrations of technology failures.

  • Details
    • Siena (on ChatGPT): ChatGPT generated strong, humorous country song lyrics (titled “Technology’s Got It Out for Me”) with verses, a chorus, and a bridge. However, it could not generate an actual audio track — only lyrics.
    • Siena (on Perplexity): When asked to generate a song based on the ChatGPT-created lyrics, Perplexity refused, claiming the lyrics were copyrighted. Siena pointed out this was incorrect since the lyrics were AI-generated. Perplexity instead created an entirely new song. This was considered a failure of the task.
    • Siena (on Gemini, paid plan): Gemini generated a full-length audio track (approximately 3 minutes) using the provided lyrics. The song was played live for attendees and was well-received. Siena downloaded both an MP3 and an MP4 with cover art.
    • Max: Praised the track and requested a copy be sent out.
    • Siena: Confirmed she would share the track with anyone who requested it.
    • Jackie Castrogiovanni: Discussed Suno, another AI music tool, noting it is reasonably accessible but has inconsistent creation limits and download restrictions on the paid plan. She also mentioned MusicGPT as inaccessible.
  • Conclusion
    • Gemini (paid plan) was the clear winner for audio song generation, producing a full MP3/MP4 track.
    • ChatGPT excelled at lyric writing but cannot generate audio.
    • Perplexity failed the task by incorrectly flagging AI-generated lyrics as copyrighted.
    • Suno remains a viable but inconsistent option for accessible AI music creation.

Differences Between AI Assistants — General Overview

Liz raised a question about why there are so many separate AI programs and whether they all operate differently.

  • Details
    • Liz: Asked whether ChatGPT, Gemini, Perplexity, and Meta AI all operate differently and noted it seemed wasteful to have so many separate systems.
    • Ali: Explained they all work on similar underlying principles but operate differently in practice.
    • Siena: Elaborated that different AIs excel at different tasks — Gemini and Perplexity are strong for research, while ChatGPT and Gemini perform better for productivity tasks. She recommended experimenting with free plans to determine which suits specific needs.
    • Siena: Noted that Meta AI (used via glasses) will be the topic of the following week’s session.
    • Palmer: Mentioned interest in Claude and the Perplexity browser, noting he primarily uses Gemini and has had some issues with ChatGPT’s voice conversation feature.
    • Siena: Mentioned Suno (music-specific AI) and Arena AI (for audio drama creation) as additional specialized tools.
  • Conclusion
    • No single AI assistant is best for all tasks; users should experiment with free plans.
    • Meta glasses and their AI integration will be covered in the next Textravaganza Afternoons session.

Gemini Pricing and Plan Clarification

Max asked about the difference between free and paid Gemini plans and whether paying for Google storage automatically includes Gemini AI features.

  • Details
    • Max: Asked whether paying for 100 GB of Gmail storage includes the paid Gemini AI plan.
    • Siena: Clarified that she pays approximately $26/month for a plan that includes 2 terabytes of Google storage and the Gemini AI Premium tier. She noted the AI Ultra plan costs $139/month, which she does not use.
    • Siena: Explained she uses Google Drive as universal cloud storage accessible from both her Windows PC and Mac, making the plan worthwhile for her workflow.
  • Conclusion
    • Paying for Google storage alone does not automatically include Gemini AI Premium.
    • The Gemini AI Premium plan (~$26/month with 2 TB storage) is separate from basic storage plans.

Mac vs. Windows Accessibility Discussion

Ali opened a broader discussion about the state of VoiceOver on macOS, frustrations with accessibility regressions, and whether switching fully to Windows is worth considering.

  • Details
    • Siena: Reported that Google Chrome on macOS has become significantly less accessible, with tasks that previously worked now broken. WordPress also no longer functions properly in Chrome on Mac and only partially works in Safari.
    • Ali: Agreed, noting VoiceOver issues are not browser-specific but are rooted in VoiceOver itself, and that accessibility has worsened with each macOS update.
    • Palmer: Noted that iPhone and iPad accessibility issues are more manageable and workable compared to macOS.
    • Siena: Said she plans to keep the Mac for specific tasks — broadcasting via Audio Hijack, Fargo (soundboard/DJ console), FaceTime, and iMessage — while using Windows for productivity, web browsing, and word processing.
    • Ali: Identified key Mac strengths he values: Audio Hijack, Loopback, Fargo, and iMessage. He noted replicating this setup on Windows would be very difficult.
    • Ali: Acknowledged he is “very forgiving” toward the Mac despite its issues, continuing to use it because of these specific strengths.
    • Lily: Suggested keeping the Mac for audio recording/broadcasting tasks where it works well, and using Windows for everything else, noting there is nothing wrong with using multiple devices for different purposes.
    • Ali: Agreed, noting everything is saved to Google Drive, making cross-device workflows seamless. He also noted Reaper works well on Windows for audio editing.
    • Ali: Highlighted that Windows with JAWS handles text selection and copy-paste from websites far more efficiently than VoiceOver on Mac, where users must use “copy last phrase” repeatedly.
    • Palmer: Confirmed he uses macOS Monterey on a Mac Neo and finds Safari faster for banking. He noted occasional VoiceOver freezes requiring a restart, and that the power button locks the screen rather than offering a shutdown option.
    • Palmer: Raised the question of whether Apple may shift focus more toward iPads and away from Mac laptops.
    • Ali: Expressed uncertainty but noted Apple’s attention is clearly not focused on improving VoiceOver on Mac.
  • Conclusion
    • Both Siena and Ali plan to retain their Macs for specific use cases (broadcasting, messaging, FaceTime) while relying more on Windows for general productivity.
    • The consensus was that using both platforms strategically — similar to choosing between AI tools — is the most practical approach.
    • VoiceOver on macOS is experiencing ongoing regressions with no clear sign of improvement.

iOS 26 VoiceOver Issue — Hey Siri Setup

Max shared a specific accessibility bug encountered on iOS 26 when attempting to re-enable Hey Siri.

  • Details
    • Max: Reported that on his iPhone 17 Pro running iOS 26, VoiceOver was unable to read the prompts displayed during the Hey Siri setup process. He tried multiple workarounds with Apple Accessibility support via screen share on a landline, including turning off screen curtain and adjusting other accessibility settings.
    • Max: The only solution was to turn VoiceOver off entirely and have the Apple support agent read the prompts aloud so he could say them to his phone.
    • Max: Expressed frustration that a core Apple feature is inaccessible to VoiceOver users, and asked how users can effectively report such issues to Apple.
    • Ali: Noted a disconnect between Apple Accessibility and the engineering team, suggesting feedback does not reliably result in fixes.
    • Siena: Empathized and confirmed the frustration, noting the screen curtain being on could sometimes cause similar issues but acknowledged Max had already ruled that out.
    • Lily: Agreed the issue is unacceptable for a native Apple feature.
  • Conclusion
    • Hey Siri setup on iOS 26 is currently inaccessible with VoiceOver enabled, requiring VoiceOver to be turned off as a workaround.
    • There is a perceived disconnect between Apple Accessibility support and the engineering team responsible for fixing such bugs.

Windows Chrome Not Launching — Palmer’s Issue

Palmer reported that Google Chrome would not open on his Windows laptop, even when searched via the Start menu.

  • Details
    • Palmer: Described that Chrome would not launch from the desktop shortcut or the Start menu search. He had been working around this by re-running the Chrome installer downloaded via Microsoft Edge each time.
    • Siena: Suggested trying to launch Chrome via the Windows Run dialog (Windows key + R, then typing “chrome”).
    • Palmer: Tried this live during the session and confirmed Chrome opened immediately.
    • Siena: Recommended removing the broken desktop shortcut and recreating it after the session.
  • Conclusion
    • The Windows Run command (“chrome”) successfully launched Chrome as an immediate workaround.
    • Palmer and Siena agreed to follow up after the session to recreate the desktop shortcut properly.

Windows Laptop Recommendations for Accessibility Users

Palmer discussed his need for a lightweight Windows laptop to replace his 8-year-old HP, suitable for accessibility work including JAWS, NVDA, and Narrator.

  • Details
    • Palmer: Explained he needs a light laptop without a number pad, with standard navigation keys (Home, Page Up, Page Down, End, arrow keys), capable of running JAWS, NVDA, and Narrator. He mentioned his current laptop’s up arrow key is failing.
    • Siena: Shared her own setup — a Lenovo Legion Go (a handheld gaming PC running full Windows 11) used as a travel computer, paired with a Bluetooth keyboard. She noted it has USB-C ports and requires a USB hub for standard USB-A devices.
    • Palmer: Asked about USB connectivity on the Legion Go; Siena confirmed it uses USB-C and a hub is needed for USB-A devices, available for under $20 on Amazon.
    • Siena: Also mentioned the ASUS ZenBook and the GPD Win Max as options she had considered, noting the GPD Win Max was out of her price range.
    • Palmer: Expressed interest in the ASUS brand. He also noted his Sony Bluetooth headphones connected successfully alongside his hearing aids via Bluetooth, which surprised him.
    • Palmer: Asked about Pixel phone availability (specifically whether the “A” series continues past the Pixel 10); Siena was unsure but thought the A series may have ended around the Pixel 10.
  • Conclusion
    • A lightweight Windows laptop without a number pad is recommended for portable accessibility work.
    • USB-C hubs (under $20) solve connectivity issues for devices with only USB-C ports.
    • The ASUS ZenBook and GPD Win Max were mentioned as options worth researching.

Screen Recording Tools for Accessibility Demonstrations

Palmer asked about accessible tools for recording screen demonstrations, having encountered issues with Loom.

  • Details
    • Palmer: Described wanting to record his screen (including screen reader output) for marketing/tutorial purposes. He tried Loom but had setup issues, possibly related to connectivity.
    • Siena: Said she is not familiar with Loom and suggested Zoom or Google Meet as accessible alternatives, noting both allow screen sharing and recording even with a single participant.
    • Siena: Explained that to record an iPhone screen via Zoom on a PC, both devices need to be connected to the same Zoom meeting, and the iPhone can be shared via a wired connection on Mac.
    • Palmer: Asked whether the webcam video needs to be active alongside screen sharing; Siena was unsure and offered to research it.
    • Siena: Noted Zoom recordings do not include built-in editing; a separate tool like Reaper (which Ali confirmed can handle video) would be needed for editing.
    • Ali: Confirmed Reaper supports video editing but noted he is not an expert in that area.
  • Conclusion
    • Zoom and Google Meet are recommended as accessible screen recording options.
    • Editing recorded content would require a separate tool such as Reaper.
    • Siena agreed to research whether webcam video is required alongside screen sharing in Zoom recordings.

AI for Music Score Reading — Liz’s Question

Liz raised the idea of using AI to help her learn piano by reading music scores note by note, given her near-total vision loss.

  • Details
    • Liz: Explained that her piano teacher writes out note names (e.g., C, C, F, E, C, D) by hand, and she uses her phone’s screen reader to read them back slowly. She asked whether AI could read a treble clef score one note at a time.
    • Siena: Expressed curiosity and suggested exploring whether AI could generate a music score from an audio file and then read it back note by note.
    • Liz: Clarified she would prefer single notes rather than chords, as chords are too complex to explain orally.
    • Siena: Noted she is an auditory learner and cannot rely on sheet music herself, but acknowledged ear training does not work for everyone.
    • Siena: Offered to research the idea and suggested Liz provide a simple song (Liz chose “Somewhere Over the Rainbow”) so Siena could attempt to provide the note letters.
  • Conclusion
    • No definitive AI solution was identified during the session, but the idea was flagged as worth exploring.
    • Siena agreed to investigate whether AI can extract and read music scores note by note.
    • Liz will share “Somewhere Over the Rainbow” with Siena as a test case.

Upcoming Sessions and Podcast Feeds

Ali and Siena outlined upcoming Textravaganza content and clarified the podcast structure.

  • Details
    • Siena: Announced a Textravaganza Extra this week focused on accessible games — Dice World and Quentin C’s Game Room.
    • Siena: Confirmed the next Textravaganza Afternoons session will cover Meta glasses and their AI features.
    • Ali: Clarified there are now three separate podcast feeds: Techstravaganza Afternoons, Techstravaganza Evenings, and Techstravaganza Extra.
    • Max: Confirmed he is busy Thursdays but asked whether the Extra session would be podcasted; Siena confirmed it would be.
  • Conclusion
    • Techstravaganza Extra (this week): Accessible games — Dice World and Quentin C’s Game Room.
    • Next Techstravaganza Afternoons: Meta glasses and AI.
    • All sessions are available as podcasts across three separate feeds.

Read more about what’s coming up on this month’s Techstravaganza Afternoons workshops at the following link: https://techstravaganza.ca/2026/09/02/whats-coming-up-on-techstravaganza-afternoons-as-of-september-2-2026/

Contact the team:

Email: support@techstravaganza.ca

Like us on Facebook: https://www.facebook.com/profile.php?id=61578990010675

Would you like to join the Techstravaganza Workshops Whatsapp Group? Fill out the form located at the following link: https://techstravaganza.ca/join-the-techstravaganza-workshops-whatsapp-group/

Follow the Show:

Tunein: http://tun.in/pxXlj

Apple Podcasts: https://podcasts.apple.com/ca/podcast/techstravaganza/id1846930800

Audible: https://podcasts.apple.com/ca/podcast/techstravaganza/id1846930800

iHeart: https://www.iheart.com/podcast/269-techstravaganza-301161434/

Spotify: https://spotify.link/vVGru9nKBXb

Comments

Leave a Reply

More posts

Discover more from Techstravaganza

Subscribe now to keep reading and get access to the full archive.

Continue reading