weekend ai reads for 2024-01-26

In this paper, we detail how AI tools have augmented teaching and learning in CS50, specifically in explaining code snippets, improving code style, and accurately responding to curricular and administrative queries on the course’s discussion forum. Additionally, we present our methodological approach, implementation details, and guidance for those considering using these tools or AI generally in education

The company added that the Practice sets feature, which uses AI to create answers and general hints, is now available in over 50 languages. Plus, educators can turn a Google Form into a practice set.

Miko Mini is the perfect companion for kids, offering meaningful interactions, realistic reactions, impressive dance moves, and engaging educational STEAM content. It is the ideal teacher and companion for kids.

To Nitta, the stronger role is to serve as an assistant to experts rather than a replacement for an expert tutor. In other words, instead of replacing, say, a therapist, he imagines that chatbots can help a human therapist summarize and organize notes from a session with a patient.

The university’s pro-vice-chancellor, Prof Rachel Thomson, said the technology could help the university achieve its sustainability strategy by reducing the need to fly in guest speakers and by facilitating international research collaborations, as well as by reducing the amount of material used by students building prototypes in engineering, design and the creative arts.

I’m sure some of them are legitimately concerned about safety, but it’s a hell of a thing how much it lines up with the strategy.

Our capacity as humans to get things right or wrong depends on our relationship to base reality, and our ability to try to solve the “inverse problem” of mapping how this reality works. LLMs don’t have that opportunity to explore and try to figure out what causes what.

Because computer vision, as it stands today, only has an economic advantage in 23% of vision tasks at the firm-level and barriers to AI-as-a-service deployments exist, there will most likely need to be a sharp reduction in cost for computer vision to replace human labor.

Even with a 50% annual cost decrease, it will take until 2026 before half of the vision tasks have a machine economic advantage and by 2042 there will still exist tasks that are exposed to computer vision, but where human labor has the advantage. At a 10% annual system cost decrease, computer vision market penetration will still be less than half of exposed task compensation by 2042.

Organizations that report very high expertise in generative AI tend to feel more positive about it—but also more pressured and threatened

Most organizations are still primarily relying on off-the-shelf generative AI solutions.

Talent, governance and risk are critical areas where generative AI preparedness is lacking

The company said it also banned applications that discouraged voting—by claiming a vote was meaningless, for example. Questions about how and where to vote would be addressed with a link to CanIVote.org, operated by the National Association of Secretaries of State.

The UK government has decided not to introduce broad copyright exceptions for text and data mining for AI systems. In addition, the reproduction of content by AI systems violates copyright law.

This analogy demonstrates that without AI, some students’ struggles will inhibit learning, like a mountain bike; while with too much reliance on and lack of understanding of AI is unpredictable and can even be harmful like a motorcycle. Ideally, AI would be used like an E bike, with the human in control.

Digital SAT Math Prep powered by Active Recall AI

100,000+ Personalized Test-Like Questions

The combination of conversational ability, embodiment, and deep user engagement shows a pathway for generalist Intelligent Social Agents to aid students in informal contexts, scaffolding their stress and mental health and even countering suicidal ideation.

But over the past year, Perplexity has evolved rapidly. It now has its own search index and has built its own LLMs based on open source models. They’ve also begun to combine their proprietary technology products.

Today, we’re excited to share our thesis for how AI development will evolve, as well as the core infrastructure components that will combine to create the modern AI stack—the new runtime architecture that will drive AI applications for the coming decade.

Four Key Design Principles for the New AI Infrastructure Stack

1. The Majority of Spend Is for Inference vs. Training

2. We Live in a Multi-Model World

3. RAG Is the Dominant Architectural Approach

4. All Developers Are Now AI Developers

To understand why AI’s responses are probabilistic, we need to understand how models generate responses, a process known as sampling (or decoding).

From its founding, OpenAI said its governing documents were available to the public. When WIRED requested copies after the company’s boardroom drama, it declined to provide them.

Over a wide range of document types, Binoculars detects over 90% of generated samples from ChatGPT (and other LLMs) at a false positive rate of 0.01%, despite not being trained on any ChatGPT data.

But far from stopping their dataset work, EleutherAI is now building an updated version of the Pile dataset, in collaboration with multiple organizations including the University of Toronto and the Allen Institute for AI, as well as independent researchers.

Multi-way parallel, machine generated content not only dominates the translations in lower resource languages; it also constitutes a large fraction of the total web content in those languages.

Code churn -- the percentage of lines that are reverted or updated less than two weeks after being authored -- is projected to double in 2024 compared to its 2021, pre-AI baseline. We further find that the percentage of “added code” and “copy/pasted code” is increasing in proportion to “updated,” “deleted,” and “moved” code.

Google is adding its “Help me write” feature to every site on the web; you just right-click on any text box anywhere, select the feature, and Google’s AI will ask you what you want to write and then generate a first draft for you. Google suggests you might use “Help me write” to write reviews and emails or RSVP to parties.

But what’s the aim of [Fictional UI]? In some cases, it’s simply to make phones seem real or narrative settings function as a whole. Rather than foreground the tech, it’s often used in the background, sometimes designed specifically for the film or show due to copyright reasons. Finder-Spyder, for example, is a bogus Google featured in countless major films.

For now, though, sit back and immerse yourself in the fictional interfaces of five futuristic shows and films—“Black Mirror”, “Her”, “Extrapolations”, “Ted Lasso” and “Red Rose”—hearing the creative process from the—thankfully for now, human—designers.