Is Brave in trouble?

Brave is my browser of choice and I have used it now for a couple of years. I used to write about it frequently, but I have neglected constantly explaining why I think it is until I read the news I will describe shortly. For the uninitiated, Brave is based on nearly the same basic code as Chrome which means it can run the same extensions. What makes Brave my choice is the system it provides for compensating content creators while maintaining security against cookies. I object to ad blockers because while they provide protection of personal data collection, view traditional ads also provide a way to compensate those who provide online content and the services we all use online. Simply put, it seems unethical to block ads while taking advantage of the work and creativity of others.

Brave does offer an option for viewing ads, but the selection is not based on personal data sent back to an ad company.

This article from TechCrunch indicates that Brave is having revenue problems and is laying off 9% of its workforce. I admit this situation does not surprise me. The ads I see are nearly all related to cryptocurrencies and security products and services. I consider these niche markets and not the type of information likely to interest most users. My guess is that Brave is struggling to attract the type of advertisers that would be making use of Google ad services. This almost seems a classic chicken and egg problem. If Brave is not regarded as a productive advertising environment, it will not attract ad money. If Brave cannot attract ad money, it will not be displaying ads a large proportion of Internet users would click.

I am not immediately impacted because Brave works just fine. It is just the company’s financial situation that concerns me. Give Brave a try, it works great and provides an approach I believe provides an ethical compromise between personal security and the legitimate needs of content and service providers.

Posted in Uncategorized | Tagged | Comments Off on Is Brave in trouble?

AI Masters Breakout

Communicating the creative potential of AI can be a challenge. Stories (examples) offer an approach if your listener understands the example you use. One story about AI that first impressed me was based on the game of chess. Most know that AI can defeat a chess master. This did not impress me. I assumed that a computer has far better memory than a human and I had always assumed that Chess Masters had so much experience that they recognized the arrangement of the pieces on the board by way of pattern recognition and could use this knowledge to call up appropriate ways to attack. I thought I knew that experienced chess players are no better than amateurs should they be asked to recall the placement of chess pieces on a board that appeared in locations that are random and not the result of a game. 

My way of understanding was challenged when told that AI had “discovered” strategies that were unique and unexpected even by experts. This would seem a form of creativity.

I understand chess at a very basic level so I take all of this on faith. I don’t really understand references to specific arrangements and named attacks and defenses. I did not really understand what a strategy undiscovered or used by experts might look like 

I do understand Breakout. Most folks of my generation and younger have played some version of Atari’s Breakout. In this simple game, a player moves a paddle at the bottom of the screen to bounce a ball upward against rows of bricks. Bricks disappear when struck by the ball. The idea is to eliminate all of the rows of bricks without missing the ball as it falls back down. 

DeepMind is an AI company. In his book “The Coming Wave”, AI pioneer Mustafa Suleyman describes when he first understood the power of the AI he was working on. The company was developing its DeepMind AI by taking on Atari games. It took a while through trial and error for the system to “understand” moving the paddle to bounce the ball to break bricks and increase the score. However, DeepMind did improve and at some point seemed to generate a strategy employed by very few human players. Instead of just keeping the ball in play and gradually working its way through the rows of bricks, DeepMind kept targeting a single row of bricks until it had created a tunnel through the rows. If you have played Breakout, you probably know what happens next. When a ball makes it through the rows of bricks it begins to bounce back and forth against the top of the screen rapidly chewing through the rows of bricks from above. By generating a tunnel through the bricks rather than chewing gradually through the rows creating a wide opening this ping-pong effect against the top before falling back continues much longer running up the score toward the end of the game at a rapid pace. Suleyman recognized that few players adopted or seemed to take this approach realizing there was something unique about how DeepMind had come to attack the game.

This story helps me understand.

BTW – Sulayman’s book is worth a read. It takes on the challenges of AI and BioTech as threats that need to be appreciated and controlled. Once a wave starts and takes form, it becomes impossible to control. With widespread adoption, there will always be bad actors who see opportunities for evil and who will be able to direct a technology toward bad ends even though positive opportunities in the technology are the most common application. 

Sulayman, M. (2023). The coming wave: Technology, power, and the twenty-first century’s greatest dilemma.  ISBN 9780593593950

Posted in Uncategorized | Tagged | Comments Off on AI Masters Breakout

Dr. Seuss Smart Notes

Smart notes represent one of the most valuable concepts I discovered in my exploration of personal knowledge management. A smart note is a note that can stand alone at a future day as providing sufficient information that you or someone else does not have to have access to the original source. For a writer, I like to describe the idea as doing some of your writing early to make later writing more efficient.

Examples can help. The following is a smart note I generated while reading a book about the reading wars. I hope you can see what I mean by containing the context necessary to be useful at a later time. I thought others might find this note interesting. The citation is included and I do this because I make use of citations in most of the things I write.

Johns, A. (2023). The science of reading: Information, media, and mind in modern America. In The Science of Reading. University of Chicago Press.

Interesting story from this author. Following the release of Flesch’s “Why Johnny Can’t Read”, there was a backlash against textbook publishers not focusing their materials on a phonics-based approach. In 1955, William Spalding the head of the education division of Houghton Mifflin turned to an old army buddy for help because he was concerned for the public image of his materials. Ted Geisel had an interest in rhyming material for children. Spalding wanted reading materials that took advantage of rhymes, would be of great interest, and were based on a limited word list. He developed a 225 list he gave to Geisel. Geisel tried for some time to come up with an approach. Finally, he took the first two words from the list “Cat” and “Hat” and started from there. Of course, Geisel is better known as Dr. Seuss and the book he produced was “The cat in the hat”. The author claims the Cat in the hat is possibly the most impactful poem of the 20th century.

Posted in Uncategorized | Tagged | Comments Off on Dr. Seuss Smart Notes

The Coming Cost of Convenience: AI Automation and Personal Services

I have reached the point that I am convinced AI tools will play a significant role in education and in our daily lives. The amounts of money companies are investing to develop and attract interest in their AI services is substantial and we have yet to recognize that we all will soon pay for these businesses in one way or another. It’s easy to forget that “free” services come with a cost, even if we don’t directly pay for them. Companies like Google and Amazon invest heavily in developing AI technology. By offering free services, they offset these investments through the data collected and opportunities to promote their other offerings. As AI progresses, developing and maintaining these complex systems will become more expensive. Companies will look to monetize AI directly to sustain continued innovation.

Some examples provide insights into what we might anticipate. Kahnmigo (the AI tutor now available to some users of the Kahn Academy) is presently priced at $9 per student per month. ChatGPT offers its best AI tool at $20 per month. I pay OpenAI for use of its API on a pay as I consume plan and pay significantly less than the $20 per month level, by using a lower-level system and paying based on the amount of use. Some tools I have recommended in the past (e.g., ChatPDF) as free up to a point and then expects payment if used extensively. So, the costs will likely vary based on the complexity of the AI system and extent of customization. Twenty dollars a month seems a reasonable estimate for individual services at this point. 

As with any new technology, finding the right balance between innovation and accessibility will be key. Through thoughtful regulation and competition within the marketplace, hopefully AI capabilities can remain affordable and available to all income levels. If companies get pricing right, AI services could enhance people’s lives in so many ways. Consumers may gripe about new fees, but ultimately may find the benefits outweigh the costs.

As an educator, I am concerned about equity issues schools will face in providing access to what I anticipate will become increasing use of AI. Consider the cost of Kahnmigo as an example. What are present investments in technology and what would be the consequence of adding another $100 per year per student fee. With all of the interest in AI, it is probably time to begin to consider the costs that will eventually become real. 

Posted in Uncategorized | Tagged , | Comments Off on The Coming Cost of Convenience: AI Automation and Personal Services

Is Big Testing driving secondary reading instruction?

I just read a Forbes article arguing that standardized testing companies have driven secondary English educators to abandon a focus on book-length literature in favor of short documents. Forbes is not exactly a scientific journal focused on careful research methods, but it is a source many would regard as credible so the article was worth a look.

Forbes describes the switch in focus I have just described as focusing on the “atomization of literature”. It acknowledges that a single book (Moby Dick was the example) might take a third or fourth of a year of instructional time, but argued that dealing with lengthy material develops important skills not satisfied by short content.

I must admit that when I first encountered this article I would have thought the focus would have been on the online reading experiences of adolescents which do mostly involve short posts. Adolescent reading habits were not mentioned by the author.

The logic for the author holding standardized testing companies responsible goes something like this. Standardized comprehension assessments rely on the understanding of short text segments – read this and then answer the following question. This is obviously true. The article then contends teachers want their students to score well on these tests and hence adjust their instruction to focus on the reading of short documents.

First, I don’t know that teachers prioritize short document comprehension or that assignments have changed. Second, I do not think it has been established that teachers make a connection between standardized reading comprehension tests and the type of tasks they assign.

I can certainly agree that different skills are required in the reading of lengthy documents or in the integration of ideas from multiple documents from those that are required to comprehend short passages. I also assume educators assign longer documents and evaluate the processing of such material through writing or other means. Whatever issues exist in relationship to the use of standardized testing and the testing industry, the complaint I have described here seems a stretch.

Posted in Uncategorized | Comments Off on Is Big Testing driving secondary reading instruction?

AI added to the Brave browser

I have reached the point at which I no longer explore the most recent AI services and concentrate instead on a limited number of AI services for the work I do. This is efficient and also allows me insights into AI capabilities at some depth. I am making an exception with this post because I am a fan of the Brave Browser and ecosystem and Brave has introduced an AI tool called Leo that is integrated into Brave.

Leo is a Llama2 large AI tool developed by Meta. Brave presently makes it available within its nightly builds so present users of the standard Brave release will not find it when this post was written. It is easy enough to download a daily build and keep this version separate from the standard version. Leo works great, but you may find other glitches if you experiment with nightly builds.

You may find accessing Leo a challenge at first. The easiest way is to look for Leo under the Settings heading and this will add Leo to the sidebar. Open the sidebar and Leo should now be present as one of the options for what appears within the sidebar

Leo will assume your interest is in interacting with the content presently appearing in your browser window. Leo will make suggestions for prompts you might apply and summarization would seem a good way to start. From that point, you can select from additional prompt suggestions or generate your own. The content generated can be copied and pasted for use in other applications.

Posted in Uncategorized | Comments Off on AI added to the Brave browser

Consensus Discovery

Consensus is an AI-powered search tool that identifies and integrates sources on a user-defined goal. Responses to queries are based on an ever-expanding database of studies now mostly from Semantic Scholar. At this time there are approximately 200,000 studies in this collection. In comparison to the training data for ChatGPT, the material used for Consensus is current. 

The search process works by identifying claims from published documents. Claims are then matched to the focus of the prompt and the most relevant are returned. Additional associated metadata such as citation count is used to label what Consensus considers the most impactful studies. 

Pricing – the free version of Consensus allows the generation of three summaries per month. The $8 per month allows unlimited queries. This is a nice combination allowing adequate room for evaluation and occasional use. For heavy use, $8 per month seems very reasonable.

Here is an example. I have been using this same topic recently because I have thoroughly explored the research and feel qualified to evaluate AI tools applied in this area. 

Prompt: What does published research conclude when comparing taking notes by hand and when using a laptop?

The service returned 20 or so studies. I would say at least 10 or so were directly relevant to the prompt. The rest addressed notetaking in different ways but not necessarily relevant to the comparison I requested.

The first images os what you first receive in response to the prompt. For each source, the claim the service identified is displayed as well as some other metadata (see red box).

Selecting one of the hits brings up additional information including a citation, the key sentence (claim), and the abstract from the source.

You cannot read the full articles unless you have a way to access them.  I have this access as a university faculty member. Before I cite research I always read the article. The abstract does not allow me to review the Method used in the research. How data are collected is important to understand when making any type of recommendation based on research and this is especially the case when studies come to different conclusions as is the case with comparisons of handwritten and keyboard notes. 

Discovery is also an important use of this type of tool. I found one reference for a study from what the system described as a rigorous source that I had not read and a few other sources that might be worth a look.

I am at the point when I need to redirect my time from exploring so many AI tools to using the tools to get work done. This field is moving so fast new approaches and ideas are constantly surfacing, but it is important not to get caught in the trap of following new opportunities that seem a little bit better than what you are now applying. The present exploration did surface a tool I will use.

Posted in Uncategorized | Tagged | Comments Off on Consensus Discovery

Three Kindles or a substack

I have no intention of abandoning the blogs I write and host, but I have been exploring what I would describe as collective hosting services – Medium and Substack. I have cross-posted a few of my blog posts to Medium for about a year and recently started to add some content to Substack

If you are unfamiliar with these services, the following information and comparisons may be helpful. 

Medium and Substack are both popular platforms for publishing and reading content, but they have some key differences:

1. Content Type:

• Medium: Medium is a platform that hosts a wide range of content, including articles, blog posts, essays, and stories. It allows writers to contribute to publications or create their own personal blogs.

• Substack: Substack is primarily focused on newsletters. It’s designed for writers to create and monetize their newsletters, which can include a mix of written content, audio, and other media.

2. Monetization:

• Medium: On Medium, writers can earn money through the Medium Partner Program, which allows members to read exclusive stories by paying a monthly fee. Writers receive a portion of the earnings based on the engagement their stories generate. A membership in Medium is $50 for the year.

• Substack: Substack offers a subscription-based model, where writers can charge a fee for access to their newsletters. If a writer wants to be paid, they designate some of their content as requiring a subscription. Readers who subscribe to read the content of individual authors pay $5 per month/$50 per year for each subscription. Subscribers get access to exclusive content, and Substack takes a percentage of the subscription revenue.

3. Ownership and Control:

• Medium: Writers retain ownership of their content but grant Medium a non-exclusive license to publish it on their platform.

• Substack: Writers retain full ownership and control of their content on Substack, which means they can export their subscriber list and move to a different platform if they choose to.

4. Community and Discovery:

• Medium: Medium has a larger built-in community, making it easier for writers to gain exposure to a broader audience. It also features curated publications that can help boost visibility.

• Substack: Substack relies more on the writer’s ability to build and nurture their own community of subscribers. Discovery on Substack often happens through word-of-mouth or external marketing efforts.

5. Design and Customization:

• Medium: Medium provides a standardized design for all posts, which maintains consistency throughout the platform.

• Substack: Substack offers more customization options, allowing writers to personalize the design and layout of their newsletters to match their branding.

Ultimately, the choice between authoring on Medium and Substack depends on the type of content you want to publish and your preferred monetization model. If you aim to create newsletters and have more control over your content and subscribers, Substack might be the better fit. On the other hand, if you want a broader audience and the ability to publish various types of content, Medium could be the more suitable option.

I do not intend to generate content exclusive to either Medium or Substack and the content I cross-post from my blogs is shared free to both outlets. As a reader, I struggle with the Substack model. I presently have one subscription to a colleague on Substack. I pay the same amount for this single subscription as the cost of a subscription to Medium. Here is a different perspective. I can purchase approximately three Kindle books for the amount I would pay for one Substack subscription. As a reader, I simply do not see the value in Substack subscriptions. I read a lot of content and I would be spending hundreds of dollars on Substack subscriptions to access the equivalent of what I read in books by established authors. To me, Medium seems a more reasonable investment.

Posted in Uncategorized | Comments Off on Three Kindles or a substack

PHET simulations

I have written about PHET simulations several times over the years. If I remember correctly, the originals were developed in Flash and with the demise of Flash, they have been reworked in HMTL5. New additions are constantly being added to the collection and they are appealing in the K12 environment because they are free.

Simulations have always interested me. They are hands-on, but many offer the important opportunity of trial and error. The opportunity to test your understanding is important because it is a proven way to modify flawed understandings of how things work. Without first activating personal theories, it is possible we can maintain flawed beliefs even when learning an alternative way of understanding.

One example those who study this issue often use is our understanding of electrical current and what is “used up” when some device uses current. For example, a light bulb uses something and we know this because we pay electric bills. It would make some sense that the current measured on one side of a light bulb would be different than on the other side. This is not the case and a simulation allows a test (so does the physical equivalent) of this idea.

Simulations are useful for other reasons, but this example of what it takes to challenge misconceptions is especially interesting.

Posted in Uncategorized | Comments Off on PHET simulations

Now on SubStack

I don’t know if this will be considered drifting over to the dark side or not. I have started adding a few of my posts to SubStack. I have no intention of requiring a subscription to view anything I cross-post to that source. The material that follows explains my motives. 

If you are a SubStack user, I have an account at https://markgrabe.substack.com

The online world is littered with my tried and abandoned accounts. That makes me an online litterer. Here is how I know. Maybe you have had this experience and recognize the problem. You learn of some new service and decide to give it a try. You apply for an account and are told your proposed username is already in use. You have been there before. You make the effort to recover your password and sure enough. You were there. Maybe there remains evidence of the last time you explored the service and maybe not. 

I maintain multiple blogs and use Facebook and Instagram. I have the Mastodon accounts and even TruthSocial. I have two photo accounts and three social bookmarking accounts. I use Notion, Obsidian, and Mem.ai. I have a paid Open.AI allocation so I can apply ChatGPT APIs. Etc.

I originally justified my exploration because I worked as an educational psychologist spending the last twenty years or so of my employment in the field researching and teaching the classroom applications of technology. This work included the generation with my wife of two textbooks. Some of what I wrote was intended to extend the content of the textbooks for adopters. Some my accounts were applied to engage with my own students. Since retirement, I can no longer use my work as an excuse. I have to admit that I just enjoy learning and creating.

Establishing an account on SubStack is very recent. In a way, missing this service for so long may seem strange. While I pay for quite a few services I use on a regular basis, but I came late to Mastodon and SubStack. I explain this avoidance for this reason. I have had at least one blog since 2002. I initially operated a blog on my own server I operated from University office. When I recognized that using University resources to host content (free) associated with our textbooks (not free) that might be seen as a conflict of interest, I began renting server space. With a personal investment in thousands of blog posts on servers I have maintained, I have been unable to write blog content through an intermediary. I see value in keeping my 20+ year investment in one location.

I recognize that I can cross post articles to both Medium and SubStack and I will probably do so. My intention is to use free posts to these services to encourage user exploration of my main collection of material. As I have tracked what brings readers to my content, I have seen a switch from RSS to search. Instead of recent posts being read at a much higher level than more dated content, readers seem more and more to arrive looking for something specific. Perhaps the communities that follow these more community oriented services may find something I write of interest and use the link to my server associated with these community accounts to read some of the other things I have written.

Posted in Uncategorized | 1 Comment