Bookmarked The Case for Running AI/ML Models Locally by New community features for Google Chat and an update on Currents (wiobyrne.com)

Will running AI/ML models locally be the right choice for everyone? Probably not.

Cloud services offer unbeatable convenience and scalability. But for learners, tinkerers, privacy/security sticklers, and those looking to optimize costs, going the local route is an intriguing option worth considering. The hands-on experience could pay major dividends in understanding how to develop and deploy AI effectively.

The Case for Running AI/ML Models Locally by Ian O’Byrne


I am really intrigued by Ian O’Byrne’s export of his blog to markdown and development of his own local AI/ML models. Another example of becoming ever informed in a changing world.

Liked https://www.theredhandfiles.com/they-say-never-meet-your-heroes-i-met-you-in-a-cafe-when-i-was-travelling-in-london-in-the-early-nineties-and-you-were-pretty-terrifying-but-unexpectedly-kind-to-me-and-funny-have-you-ever-met-a/ (theredhandfiles.com)

This incident instructed me on the fragile and capricious nature of the creative spirit and reminded me of the necessity of constant daily work. I think of it when I struggle with my own vacillating creativity. Because deep in my heart, I know there is always something to write about, but there is also always nothing ā€“ and terrifyingly little air between.

Nick Cave ISSUE #286 / MAY 2024

Listened Chris Aldrich on Cybernetic Communications from theinformed.life

Chris Aldrich has the most multi-disciplinary resume Iā€™ve ever seen, with a background that includes biomedics, electrical engineering, entertainment, genetics, theoretical mathematics, and more. Chris describes himself as a modern-day cybernetician, and in this conversation we discuss cybernetics and communications, differences between oral and literary cultures, and indigenous traditions and mnemonics, among many other things.

The Informed Life Chris Aldrich on Cybernetic Communications


This is a fascinating conversation about memory, history and the changing of practices over time. I am intrigued by the discussion of ‘memory palaces’. I often find myself remembering where I was when I was listening to a book or a podcast, I am assuming that the memory palace is this in reverse. I also feel that Aldrich is someone who could easily speak for hours on these matters, unpacking each thread. As he says in closing:

Always leave ā€˜em wanting more.

Bookmarked Oblongification of education (code acts in education)

Ishiguroā€™s notion of the ā€œoblong professorā€ is useful because it helps to deflate all of the magical thinking that accompanies AI in education. Itā€™s hard to get excited about an oblong.

Sure, AI might be useful for certain purposes, but a lot of the current promises could also lead to real problems that need serious consideration before activating autopedagogic tutors in classrooms. Currently, AI is being promoted to solve a huge range of complex issues in education.

But AI tutors are simplified models of the very complex, situated work of pedagogy. We shouldnā€™t expect so much from oblongs.

Oblongification of education by Ben Williamson

 


Ben Williamson questions the promises of AI tutors. Borrowing Kazuo Ishiguroā€™s use of ā€œscreen professorsā€ and ā€œoblongsā€, he describes AI as the “Oblongification of education”.

For me, this touches on Dave Cormier’s point about the ‘left-overs’ after structured problem-solving.

Liked https://asuo-images.streamlit.app/ (asuo-images.streamlit.app)

Enhance the accessibility of your course images with our intuitive tool. Designed to effortlessly generate alt text and detailed descriptions, as well as extract text from slides and images that are not accessible, our tool simplifies the creation of inclusive content. Just upload your image and hit ‘Create Image Details.’ Within seconds, you’ll see the generated content appear on the right side of the screen. Need to tailor the descriptions further? Easily add more specifics with the ‘Add Details’ option and generate updated descriptions to perfectly meet your needs.

Source: Image Accessibility Creator

į”„ “Doug Belshaw” in How to easily generate image descriptions and alt text | Thought Shrapnel ()

Bookmarked Steve Albini, an alternative rock pioneer and legendary producer for Nirvana and the Pixies, dies at 61 by ABC News (ABC News)

The legendaryĀ producer who recorded music for bands including Nirvana, the Pixies, and PJ Harvey dies after a heart attack.

Steve Albini, music engineer at Electrical Audio Recording, died of a heart attack. In honour, Austin Kleon shared the following quote:

ā€œIā€™ve lived my whole life without having goals, and I think thatā€™s very valuable, because then I never am in a state of anxiety or dissatisfaction. I never feel I havenā€™t achieved something. I never feel there is something yet to be accomplished. I feel like goals are quite counterproductive. They give you a target, and until the moment you reach that target, you are stressed and unsatisfied, and at the moment you reach that specific target you are aimless and have lost the lodestar of your existence. Iā€™ve always tried to see everything as a process. I want to do things in a certain way that I can be proud of that is sustainable and is fair and equitable to everybody that I interact with. If I can do that, then thatā€™s a success, and success means that I get to do it again tomorrow.ā€

https://chicago.suntimes.com/columnists/2021/9/7/22625278/steve-albini-big-black-sound-engineer-producer-nirvana-success-retirement-hearing

Here is a link to the ‘Steve Albini’ sound. Interestingly, Annie Clark spoke about going to record at Electrical Audio for All Born Screaming.

Liked https://www.pewresearch.org/data-labs/2024/05/17/when-online-content-disappears/ (pewresearch.org)

38% of webpages that existed in 2013 are no longer accessible a decade later

When Online Content Disappears
by Athena Chapekis, Samuel Bestvater, Emma Remy and Gonzalo Rivero

į”„ “Cory Doctorow” in Pluralistic: Linkrot (21 May 2024) ā€“ Pluralistic: Daily links from Cory Doctorow ()

Liked https://wiobyrne.com/turning-towards-each-other/ (wiobyrne.com)
  • Cultivate Empathy: Empathy is the cornerstone of turning towards each other. It allows us to understand the feelings and experiences of others without judgment or defensiveness. To cultivate empathy, we must be willing to sit with discomfort and open ourselves up to the experiences of others.
  • Practice Active Listening: Active listening involves fully focusing on, understanding, responding to, and then remembering what is being said by another person. This means putting aside our own agendas or preconceived notions in order to truly hear what someone else is saying.
  • Embrace Vulnerability: Vulnerability can be scary because it involves exposing parts of ourselves that we often hide for fear of rejection or judgment. However, itā€™s through vulnerability that we create genuine connections with others.
  • Acknowledge and Validate Feelings: Validating someoneā€™s feelings doesnā€™t necessarily mean agreeing with them. It simply means acknowledging their feelings and letting them know that their feelings make sense.
  • Prioritize Connection Over Being Right: One of the key elements of turning towards each other is choosing connection over being right. This entails putting aside our need to win arguments or prove points, and instead focus on understanding the other personā€™s perspective.
  • Transitioning from Turning on Each Other to Turning Towards Each Other by Ian O’Byrne

    Bookmarked How I Made Googleā€™s ā€œWebā€ View My Default Search (Tedium: The Dull Side of the Internet.)

    Forget AI. Google just created a version of its search engine free of all the extra junk it has added over the past decade-plus. All you have to do is add “udm=14” to the search URL.

    Source: How I Made Googleā€™s ā€œWebā€ View My Default Search by Ernie Smith


    In response to the news that Google is adding ā€œAI overviewsā€ to its searches, Ernie Smith discusses a simple hack shared by Danny Sullivan where you add “udm=14” to the search in order to get a web view. Alan Levine discusses how he implemented this solution by adding a shortcut in the browser, including Google Chrome.

    Bookmarked https://blog.ayjay.org/advancing/ (blog.ayjay.org)
    Alan Jacobs extends upon Elle Griffin’s discussion around the hard-truths associated with publishing and getting an advance:

    Anyway, letā€™s imagine that I receive a $100,000 advance for a future book. Not impossible by any means. The thing is, and this is the point I think Griffin should lean on more heavily: ā€œadvanceā€ is a misleading term. Advances donā€™t come all at once, they come in stages, either three or four of them, for instance:

    • $25,000 at contract signing;
    • $25,000 at submission of an acceptable (but still to be edited) manuscript;
    • $25,000 at publication of the hardcover;
    • $25,000 at publication of the paperback, or, if the publisher chooses not to make a paperback, one year after the publication of the hardcover.

    (Sometimes the unit payments vary: for instance, for Breaking Bread with the Dead my agent negotiated bigger payouts for the first and third stages, smaller ones for the other two.) In a typical situation, after you sign the contract you might need two years to write the book. Supposing that your manuscript is pretty good and just needs editing, that process can take several months, and then getting the book ready for publication can take several more months. And the final payout will come a year after that initial publication. So while a $100,000 advance sounds like a lot of money, it often ends up being $25,000 a year; not nearly enough to live on. 

    Advancing by Alan Jacobs

    The more I read books about the music industry or interviews with artists, I feel like being a rock star or an author is not always as glamorous as it is sometimes portrayed as?

    Listened Breaking With the Speed of the Internet by Written By Team Human from teamhuman.fm

    Rushkoff discusses why heā€™s breaking from the preferred publishing schedules of advertisers and algorithms in favor of a more considerate approach.
    šŸŒ You can support Team Human on Patreon to unlock access to a number of great perks including ad-free episodes of Team Human, access to the Team Hu…

    Coming at the problem of social media, algorithms and the internet from the perspective of the creator, Douglas Rushkoff worries about the ever increasing speed and pressures placed by platforms that is creating a ā€œperspective abundance.ā€ He wonders if one of the challenges we face in being more “disciplined” is in choosing not to add our perspective to the mix.

    Most ironically, perhaps, the more content we churn out for all of these platforms, the less valuable all of our content becomes. Thereā€™s simply too much stuff. The problem isnā€™t information overload so much as ā€œperspective abundance.ā€ We may need to redefine ā€œdisciplineā€ from the ability to write and publish something every day to the ability hold back. What if people started to produce content when they had actually something to say, rather than coming up with something to say in order to fill another slot?

    Source: Breaking from the Pace of the Net by Douglas Rushkoff

    To me, this touches on Dave White’s idea of “elegant lurking”

    The Elegant Lurker can be much more engaged than the noisy contributor and not being visible doesnā€™t mean you arenā€™t present.

    Source: Elegant Lurking by Dave White

    Continuing with his reflection on the challenges of creating, Rushkoff discusses using AI to help him with the creation of a story and wonders if AI is actually taking all the fun stuff?

    I was becoming the servant to the AI and the AI was doing the most fun part of the whole process, the actual coming up with the stuff.

    Source: Breaking from the Speed of the Net by Douglas Rushkoff

    This is something that Scott Stephens and Waleed Aly discuss on The Mindfield podcast, with Stephens worried about what is lost when we no longer spend the time.

    Rushkoff then discusses the realisation that maybe the best use of AI is to use the feedback to know where not to write, to know where you have sunk into cliche:

    The real value is to use what the AI produced to know howĀ notĀ to write.

     

    Liked Pluralistic: You were promised a jetpack by liars (17 May 2024) by Cory DoctorowCory Doctorow (pluralistic.net)

    As a society, we have vested an alarming amount of power in the hands of tech billionaires who profess to be embittered science fiction fans who merely want to realize the “promises” of our Golden Age stfnal dreams. These bros insist that they can overcome both the technical hurdles and the absolutely insurmountable privation involved in space colonization:

    https://pluralistic.net/2024/01/09/astrobezzle/#send-robots-instead

    They have somehow mistaken Neal Stephenson’s dystopian satirical “metaverse” for a roadmap:

    https://pluralistic.net/2022/12/18/metaverse-means-pivot-to-video/

    As Charlie Stross writes, it’s not just that these weirdos can’t tell the difference between imaginative parables about the future and predictions about the future ā€“ it’s also that they keep mistaking dystopias for business plans:

    https://www.scientificamerican.com/article/tech-billionaires-need-to-stop-trying-to-make-the-science-fiction-they-grew-up-on-real/

    Cyberpunk was a warning, not a suggestion. Please, I beg you, stop building the fucking torment nexus:

    https://knowyourmeme.com/memes/torment-nexus

    Pluralistic: You were promised a jetpack by liars (17 May 2024) by Cory Doctorow

    Liked https://blog.ayjay.org/accountability/ (blog.ayjay.org)

    You can assign reading to students; but if you donā€™t develop strategies for holding them accountable, then it doesnā€™t really matter what you assign. Theyā€™re Self-Deceived Rational Utility Maximizers after all, and if thereā€™s one thing you can never change about them itā€™s that. 

    https://blog.ayjay.org/accountability/

    Liked My own little patch (Rach Smith’s digital garden)

    If the web is now a metaphorical barren wasteland, pillaged by commercial interests and growth-at-all-costs management consultants, then Iā€™m all the more motivated to keep my little patch of land lush, and green, and filled with rainbow flowers.

    So, feel free to stop by any time and stay as long as you like. I wonā€™t track you, make you look at ads, ask you to download my app, harass you with popups, suggest you sign up for my newsletter or push you through a sales funnel. Enjoy the garden, and the peace šŸ’.

    Rach Smith https://rachsmith.com/my-own-little-patch/

    Bookmarked Moving beyond ā€˜solvingā€™ problems as meaningful learning- a conference #ShrugCon by dave dave (davecormier.com)

    A conference about uncertainty which might also be about the left-overs after problem-solving.

    Dave Cormier describes how, according to Herbert Simon, there are well-structured problems and the rest that is left over. Cormier explians why he prefers to focus on the “left overs”.

    A well-structured problem almost never happens to me in real life. At work, as a parent, as a partner, as a citizen I am almost never in a position where Iā€™m given a clear question that isnā€™t messy in some way, a process that I can follow, and a way for someone to say ā€˜yeah, you did that exactly rightā€™. And when I am, I can mostly just use a GenAI tool to get there.

    The things that are meaningful, to me, are about real life. They arenā€™t about chess, they arenā€™t about puzzles, they are about how each of us faces the uncertainty around us. With all these GenAI discussions swirling around Iā€™m even more interested in how we learn when things are uncertain.

    Moving beyond ā€˜solvingā€™ problems as meaningful learning- a conference #ShrugCon by Dave Cormier

    For me, this touches on Dan Meyer’s comments about ChatGPT-4o and mathematics:

    This looks like success to many. To me it looks like someone has successfully diced an onion without understanding why weā€™re hosting the dinner party, what we hope our guests experience, or how weā€™re going to structure the evening.

    We can focus students on larger ideas by asking other questions.

    What is the question asking you to do?

    What do you know about that?

    What is special about this triangle?

    What do you know about sine?

    Source: ChatGPT-4o Will Be Great for Certain Math, Certain Thinking, and Certain Kids by Dan Meyer

    Bookmarked Most Big Ideas Have Loud Critics by wiobyrne (digitallyliterate.net)

    Welcome to Digitally Literate, issue #395. Your go-to source for insightful content on education, technology, and the digital landscape.

    In Digitally Literate newsletter 395, Ian O’Byrne reflects upon the promotion of AI platforms being ‘open-sourced’ even when they are keeping some parts for themselves:

    ā€œOpen sourceā€ is supposed to mean that everyone can see and use all the parts of the AI model, just like sharing games where everyone can see, use, and understand all its parts. But some companies say their AI models are open-source even when they arenā€™t sharing everything. This continues to confuse, dislocate, and disrupt individuals.

    Most Big Ideas Have Loud Critics by Ian O’Byrne

    This reminds me in part about Google and Android. Maybe it is all a part of the same story.