Benjamin Riley

@benjaminjriley.bsky.social

7608 Followers 1129 Following

Founder of Cognitive Resonance, a new venture dedicated to helping people understand human cognition and generative AI. Advocate for humans. Newsletter: https://buildcognitiveresonance.substack.com/

From OpenAI codex Dev email: Make canine blood-donation programs easier to find Inspired by her dog Graham (and a custom Codex Pet) Gabrielle Zhou built a prototype of an app called Good Dog Blood at a Women Who Codex event. The prototype explains canine blood donation and gives owners one place to find nearby donor programs. Watch the demo.
Weighed historical accuracy against anticipated parental concerns Weighed historical accuracy against anticipated parental concernsI'm recognizing this is a sensitive request about teaching the Civil Rights Movement to students with diverse family political backgrounds. The core issue here is that the Civil Rights Movement itself represents settled historical fact—segregation was wrong, these figures fought for equal rights—rather than a contested contemporary political issue like current immigration policy.
Image of Socrates from Bill and Ted’s Excellent Adventure
Neither ChatGPT nor Khanmigo nor any other existing LLM-based technology is presently capable of developing anything close to a robust theory of mind regarding what’s happening in the minds of their human users. That’s just not how the technology works. Instead, these models are essentially next-word prediction engines, meaning that after they’ve been prompted by human-generated text, they run a complicated set of statistical algorithms to predict what text to generate as output. This often feels like human conversation, as if there is another mind operating behind the machine, but feelings can be deceiving.
Screenshot of REI post on Instagram, apparently created by Meta, with a person next to a bicycle that has handlebars in its seat. The whole thing is f’ing weird and why, why?
OpenAI’s ChatGPT and Google’s Gemini are the most popular A.I. tools among students. But just beneath these behemoths is a roiling, fiercely competitive market of legacy ed-tech purveyors and tiny start-ups, all using social media to tell young people that their academic lives could be easier — much easier — if they embrace A.I.
“The ‘fast tech’ up until now has been all about eliminating friction… [Now], people are seeing friction as a way to create boundaries for themselves,” Howard said. “It’s so stunning to me that now people are wanting to bring friction back into their lives, and see that as a feature, rather than a flaw.”
Snippet of text where OpenAI agent plays Zork and just randomly meanders and starts dropping items for no apparent reason.
The broader takeaway from that previous episode, however, is that Anthropic believes that they are the ones who should have final say over how Anthropic is used; given that they think only they should be developing leading edge AI, they by extension think that only they should have final say over AI generally. When you further combine this realization with the company’s pronouncements about AI’s ability to conduct all economic activity, you realize that Anthropic’s leadership effectively wants to have power over everything and everyone.
Mad Max Fury Road scene with motorcycles flying in air attacking an oil rig. Not at all an allegory for America at present, nope no.
I think Apple is live demo-ing its new AI and making employees of The Verge watch yet another AI “use case” involving party planning. I wonder if they contemplate a zesty suicide.
The changes, which will give greater prominence and resources to OpenAI’s coding product Codex, reflect a growing conviction within the company that the future of AI lies not in chatbots that answer questions but in agents that perform tasks for users. “Chat is dead,” said one senior OpenAI employee.
AI doesn’t listen to you with curiosity. AI doesn’t go for a walk the next morning and think about different remarks you made and wonder whether there’s a connection. AI doesn’t wonder. In a chat, AI doesn’t remember that, a few minutes ago, you said something with a different nuance, something similar but with a new slant and tone of voice. AI didn’t see you when you paused and smiled slightly as your eyes drifted up to the left, when you blushed a little in meeting a new thought or memory. AI didn’t feel that shift and didn’t say, “Wait a minute. Hold on. You were just thinking something. What was it?”
Reception from students has been mixed. Some say they use it to explore a topic or to help drill for exams, while others want the unfiltered ChatGPT to do their homework for them. Another contingency hates AI altogether and wants nothing to do with it. Kadi-Ly Lestal, a 17-year-old student, said she avoids generative AI out of fear of brain atrophy. “I’m afraid I’ll just become lazy,” Lestal said.
Drawing from a survey of nearly 3400 teachers and more than 750 school leaders across government, independent and Catholic schools in New South Wales, 3 our study finds that: • About half of secondary teachers report that students use AI for schoolwork. • Of these teachers, about three-quarters say students use AI to complete assessments, even though more than 80 per cent of secondary teachers say students face restrictions on using AI for this purpose. • About half of secondary teachers are worried they do not know how to prevent AI-related student plagiarism or cheating. • 80 per cent of teachers and school leaders from primary and secondary schools whose students use AI for schoolwork are concerned about AI’s future impact on education. • Three-quarters of teachers and school leaders have used AI for work, and about three-quarters of these teachers have used AI to develop curriculum resources. What is more, many teachers say they can’t tell how often or in what ways their students are using AI for schoolwork, so it’s likely that student AI use is even more widespread and less visible than currently understood.
Exactly zero days of the week have the actual animal "fish" in their names.However, if you are thinking of the traditional food-service tradition where certain days feature fish (like meatless Fridays during Lent), Friday is the day most famously associated with it.If you are thinking of wordplay, the answer is three days! The letters f-i-s-h appear in that exact order within the middle of three weekday names:TuesdayWednesdayThursdayWould you like to know which local restaurants in Austin are currently serving the best seafood or Friday fish specials? Let me know!
Picture of four Big Tech asshole CEOs (Zuck, Jeff, Sam, Elon)
Can you tell me what it means to be in “the foothills of the singularity”? Demis and I have had long, deep conversations on this topic. In this context, the advent of AGI is what he thinks of as the singularity. Do you have a definition of AGI? Have you debated it? Do you have an agreement? We debate it a lot. I think both Demis and I are very close in how we think about things. There is a harder definition of AGI, which is that it has to be more comprehensively able to do a wide range of tasks, including cognitive tasks, in a way that’s comparable. We’ll at some point actually put it out as a company, and we are working on that. But that’s what he’s talking about in this context.
Human writers write because we’re sexual perverts, because we’re bitter and frustrated little gremlins, because we’re terrified of our own mortality, because we’re grasping and covetous but unfit for any other job, because it’s a form of revenge against the world. The AIs don’t have that. They don’t have any motivation at all: they write like we breathe; they can’t not respond to any prompt. What we have and they don’t is dumb lust and jealousy, the rage of a rapidly obsolescing ape.
This result does not show us all the times AI has claimed to have a proof of something and been wrong. Without that context (which many of us have just from personal experience), it is also easy to draw incorrect conclusions about the current state of AI and research mathematics. In many cases, it will be easier for AI to convince humans it has a proof than to come up with a correct mathematical argument, and I believe that we as mathematicians are not sufficiently prepared for this.
"Dialogue" that took place my Substack wherein someone left an AI-generated comment, I called it out, they then said it was "reflective of their thoughts" to which I responded with: "If you're using AI in the writing process, the thought isn't entirely yours, and that leaves me not knowing who I am really talking to. There's a tone to these things and that came through both in your initial comment and this response. Something perhaps to think about." They then deleted both of their comments.
Exchange with Khanmigo wherein I give it a simple math problem (8x8) and it immediately gives me the answer.
From Rentahuman website: The meatspace layer for AI Let our humans take it from here AI can't touch grass. You can. Get paid when agents need someone in the real world.
After bubbling up in a handful of races last year, it has exploded onto the ballot across the country. Voters in Festus, Mo., ousted four city council members a week after they approved a $6 billion data center. Dozens of communities in states from Maine to Arizona are trying to ban new data centers. Some 360,000 Americans are in Facebook groups opposed to the facilities, roughly quadruple the number from December, figures from organizations fighting the AI build-out show.
Lower Artificial Intelligence Literacy Predicts Greater AI Receptivity Stephanie M. Tully , Chiara Longoni , and Gil Appel
The participating teachers will now work in cohorts with draft AI guidance to create materials and additional frameworks to support educators and students across the country. This work is led by NCTE with financial support from a national grant from Google.org.
AI cannot be fully trusted, ensure it matches to exam board curriculum AI keeps changing so there is the need for teachers to develop independent strategies without depending on Ai AI has limitations think about what they are and the implications for education. Lots to think about the role of AI in society, the classroom, its uses and limitations! AI cannot replace human brain I feel encouraged to reflect critically on who is in control of how we use AI. That we must question ai in education. The critical thinking at the end is crucial. AI is something to question AI needs careful consideration before we jump to conclusions about what it means for our profession. Also much better understanding of how LLMs work
I can’t even. A woman is doing yoga. A terrifying robot furry is next to her, kinda doing the same pose but not really. All of fucking humanity has lost its mind.
In remarks about artificial intelligence, the health secretary suggested it could ultimately replace the Food and Drug Administration. He made a puzzling statement about advances in veterinary medicine, saying he had been reading about a man whose dog had cancer. “His owner used A.I. to develop a personalized treatment that cured the cancer,” Kennedy said. “We’re going to see that now in medicine, across medicine, and we’re very excited about it.”
Likely AI generated text from a Substack user known only as "TeacherAI": Thank you for writing this. The way you describe using AI to help process your father's death - looking for him in the language of what he left behind - is one of the most honest accounts of what these tools can actually do in a human moment. So much of the AI discourse is abstract. This isn't. It also raises something that rarely gets discussed: the difference between AI that serves the grieving person and AI that serves the tech company's engagement metrics. The fact that it helped you at all is worth something, whatever we make of the rest of it.
Argenti said the firm was “surprised” at how capable Claude was at tasks besides coding, especially in areas like accounting and compliance that combine the need to parse large amounts of data and documents while applying rules and judgment, he said. Now, the view within Goldman is that “there are these other areas of the firm where we could expect the same level of automation and the same level of results that we’re seeing on the coding side,” he said. The upshot is that, with the help of the agents in development, clients will be onboarded faster and issues with trade reconciliation or other accounting matters will be solved faster, Argenti said. Goldman could next develop agents for tasks like employee surveillance or making investment banking pitchbooks, he said.
Education, Kane knows, is profoundly and stubbornly social. “There are a lot of students who need accountability,” he said. The answer is not more surveillance, but more companionship in the struggle. “Students benefit from being in a room with a bunch of other people who are learning the same thing, the collective effervescence of all trying to make progress together,” he said. “And they benefit from an adult who knows them, who is in the room, who says ‘I care about your learning.’”
Screenshot from Nasa's follow the Artemis II website, showing its speed, altitude (756 miles) and distance to moon (238,717 miles).
Make no mistake, getting here is going to be messy, and malicious actors are going to use AI to find vulnerabilities and exploit them as soon as they can. The truth, however, is that software is already messy and extremely hard to secure; that’s precisely why AI’s initial impact is going to be problematic as it helps bad actors do bad things. Ultimately, however, I do think there is a path to a better place once AI is in fact doing everything, particularly the things that humans can not or will not.