Plotting

The Time Sam Altman Asked for a Countersurveillance Audit of OpenAI

WIRED

Dario Amodei's AI safety contingent was growing disquieted with some of Sam Altman's behaviors. Shortly after OpenAI's Microsoft deal was inked in 2019, several of them were stunned to discover the extent of the promises that Altman had made to Microsoft for which technologies it would get access to in return for its investment. The terms of the deal didn't align with what they had understood from Altman. If AI safety issues actually arose in OpenAI's models, they worried, those commitments would make it far more difficult, if not impossible, to prevent the models' deployment. Amodei's contingent began to have serious doubts about Altman's honesty.


AI has entered the therapy session -- and its recording you

Mashable

As generative artificial intelligence becomes embedded in people's everyday lives, one emerging aspect of its use in mental health care is raising complicated questions about professional ethics and patient privacy. A number of companies, like Upheal, Blueprint, and Heidi Health, have begun offering AI-powered tools designed to make therapists more efficient at documenting sessions and completing administrative paperwork. Providers are typically required to record the entirety of their session with a client. While it's ethical for therapists to record these conversations under certain circumstances, it's rarely done outside of professional training and forensic work. Note-taking tools, or "scribes," use AI to analyze the content of a client's conversation with their therapist in order to generate documentation that therapists must submit for a variety of reasons, including for insurance payments and potential quality audits.


Sony is giving away a free TVs with select purchases. Heres how to get yours.

Mashable

SAVE OVER 500: Until June 8, Sony is giving away TVs for free with select purchases. Learn more about how to qualify here. Memorial Day is just around the corner, which means deals are incoming. You'll find sales on everything from laptops to clothing, but we think we've just found the deal of all deals. Until June 8, if you buy a select TV from Sony, you'll get another one completely free. Yes, you read that right.


Appendix: Symbolic Distillation for Learned TCP Congestion Control

Neural Information Processing Systems

Y[n.total_condition] Slices DRL behavior dataset, one could then apply 8: IF Entropy(Y It is an indicator of the population of genetic programs' performances. The fitness metric driving our evolution is simply the MSE between the predicted action and the "expert" action (teacher model's action). These individuals are mutated before proceeding to following evolution rounds. We specifically follow 5 different evolution schemes, either one picked stochastically. They are: Crossover: Requires a parent and a donor from two different evolution tournamets.


Symbolic Distillation for Learned TCP Congestion Control

Neural Information Processing Systems

Recent advances in TCP congestion control (CC) have achieved tremendous success with deep reinforcement learning (RL) approaches, which use feedforward neural networks (NN) to learn complex environment conditions and make better decisions. However, such "black-box" policies lack interpretability and reliability, and often, they need to operate outside the traditional TCP datapath due to the use of complex NNs. This paper proposes a novel two-stage solution to achieve the best of both worlds: first to train a deep RL agent, then distill its (over-)parameterized NN policy into white-box, light-weight rules in the form of symbolic expressions that are much easier to understand and to implement in constrained environments. At the core of our proposal is a novel symbolic branching algorithm that enables the rule to be aware of the context in terms of various network conditions, eventually converting the NN policy into a symbolic tree. The distilled symbolic rules preserve and often improve performance over state-of-the-art NN policies while being faster and simpler than a standard neural network. We validate the performance of our distilled symbolic rules on both simulation and emulation environments.


Google's new AI shopping tool just changed the way we shop online - here's why

ZDNet

In recent years, Google Search's shopping features have evolved to make Search a one-stop shop for consumers searching for specific products, deals, and retailers. Shoppers on a budget can scour Search's Shopping tab during major sale events to see which retailer offered the best deal and where. But often, consumers miss out on a product's most productive discount, paying more later because they don't want to wait again. During this year's Google I/O developer conference, Google aims to solve this problem with AI. Shopping in Google's new AI Mode integrates Gemini's capabilities into Google's existing online shopping features, allowing consumers to use conversational phrases to find the perfect product.


I tried Samsung's Project Moohan XR headset at I/O 2025 - and couldn't help but smile

ZDNet

Putting on Project Moohan, an upcoming XR headset developed by Google, Samsung, and Qualcomm, for the first time felt strangely familiar. From twisting the head strap knob on the back to slipping the standalone battery pack into my pants pocket, my mind was transported back to February of 2024, when I tried on the Apple Vision Pro during launch day. Also: Xreal's Project Aura are the Google smart glasses we've all been waiting for Only this time, the headset was powered by Android XR, Google's newest operating system built around Gemini, the same AI model that dominated the Google I/O headlines this week. The difference in software was immediately noticeable, from the starting home grid of Google apps like Photos, Maps, and YouTube (which VisionOS still lacks) to prompting for Gemini instead of Siri with a long press of the headset's multifunctional key. While my demo with Project Moohan lasted only about ten minutes, it gave me a fundamental understanding of how it's challenging Apple's Vision Pro and how Google, Samsung, and Qualcomm plan to convince the masses that the future of spatial computing does, in fact, live in a bulkier, space helmet-like device. For starters, there's no denying that the industrial designers of Project Moohan drew some inspiration from the Apple Vision Pro.


Chaos, Extremism and Optimism: Volume Analysis of Learning in Games

Neural Information Processing Systems

We perform volume analysis of Multiplicative Weights Updates (MWU) and its optimistic variant (OMWU) in zero-sum as well as coordination games. Our analysis provides new insights into these game/dynamical systems, which seem hard to achieve via the classical techniques within Computer Science and ML. First, we examine these dynamics not in their original space (simplex of actions) but in a dual space (aggregate payoffs of actions). Second, we explore how the volume of a set of initial conditions evolves over time when it is pushed forward according to the algorithm. This is reminiscent of approaches in evolutionary game theory where replicator dynamics, the continuous-time analogue of MWU, is known to preserve volume in all games. Interestingly, when we examine discrete-time dynamics, the choices of the game and the algorithm both play a critical role. So whereas MWU expands volume in zero-sum games and is thus Lyapunov chaotic, we show that OMWU contracts volume, providing an alternative understanding for its known convergent behavior. Yet, we also prove a no-free-lunch type of theorem, in the sense that when examining coordination games the roles are reversed. Using these tools, we prove two novel, rather negative properties of MWU in zero-sum games.


66de6afdfb5fb3c21d0e3b5c3226bf00-AuthorFeedback.pdf

Neural Information Processing Systems

We address your questions/suggestions below. If the paper is accepted to NeurIPS 2020, we will have one extra page. There are three players, each with two strategies Head and Tail. Indeed, we create more figures and combine them as a video. We also run with Optimistic MWU.


Is Google's 250-per-month AI subscription plan worth it? Here's what's included

ZDNet

If you're one of the 8% of Americans who say they're willing to pay for AI, Google has a deal for you -- a 250 per month AI subscription. The company unveiled Google AI Ultra today, a plan with the biggest usage limits for Google's suite of AI tools and access to the highest versions of those tools. Google AI Ultra is intended for filmmakers, developers, and creative professionals and gives users access to tools like Veo, Imagen, Whisk, NotebookLM, and a new tool called Flow. Also: Google's popular AI tool gets its own Android app - how to use NotebookLM on your phone Subscribers also get a massive expansion in storage across Google platforms, plus YouTube Premium ( 13.99 per month on its own). Here's a full breakdown of what the new plan includes: Google said the current AI Premium plan is also getting an upgrade -- to Gemini AP Pro.