Tootfinder

Snowpark: Performant, Secure, User-Friendly Data Engineering and AI/ML Next To Your Data
Brandon Baker, Elliott Brossard, Chenwei Xie, Zihao Ye, Deen Liu, Yijun Xie, Arthur Zwiegincew, Nitya Kumar Sharma, Gaurav Jain, Eugene Retunsky, Mike Halcrow, Derek Denny-Brown, Istvan Cseri, Tyler Akidau, Yuxiong He
https://arxiv.org/abs/2508.05904…

Snowpark: Performant, Secure, User-Friendly Data Engineering and AI/ML Next To Your Data
Snowflake revolutionized data analytics with an elastic architecture that decouples compute and storage, enabling scalable solutions supporting data architectures like data lake, data warehouse, data lakehouse, and data mesh. Building on this foundation, Snowflake has advanced its AI Data Cloud vision by introducing Snowpark, a managed turnkey solution that supports data engineering and AI and ML workloads using Python and other programming languages. This paper outlines Snowpark's design obj…

DINA: A Dual Defense Framework Against Internal Noise and External Attacks in Natural Language Processing
Ko-Wei Chuang, Hen-Hsen Huang, Tsai-Yen Li
https://arxiv.org/abs/2508.05671

DINA: A Dual Defense Framework Against Internal Noise and External Attacks in Natural Language Processing
As large language models (LLMs) and generative AI become increasingly integrated into customer service and moderation applications, adversarial threats emerge from both external manipulations and internal label corruption. In this work, we identify and systematically address these dual adversarial threats by introducing DINA (Dual Defense Against Internal Noise and Adversarial Attacks), a novel unified framework tailored specifically for NLP. Our approach adapts advanced noisy-label learning me…

Cerebras announces the $50/month Code Pro and the $200/month Code Max plans, offering users access to Qwen3-Coder at speeds of up to 2,000 tokens per second (Daniel Kim/Cerebras)
https://www.cerebras.ai/blog/introducing-cerebras-code

Cerebras
Cerebras is the go-to platform for fast and effortless AI training. Learn more at cerebras.ai.

Cloudflare will now block AI crawlers by default
The internet architecture provider will also let some publishers make known AI scrapers pay to crawl their sites
https://www.theverge.com/news/695501/cloudflare-block-ai-crawlers-default

Cloudflare will now block AI crawlers by default
Cloudflare has announced that it will now block AI web crawlers by default for new customers. It’s also introducing a new “Pay Per Crawl” fee that will let some publishers make AI companies pay to scrape their content.

Hey @…, how’s this for a CC signal for AI, you clowns?
🖕
https://creativecommons.org/2025/06/25

Introducing CC Signals: A New Social Contract for the Age of AI - Creative Commons
CC Signals © 2025 by Creative Commons is licensed under CC BY 4.0 Creative Commons (CC) today announces the public kickoff of the CC signals project, a new preference signals framework designed to increase reciprocity and sustain a creative commons in the age of AI. The development of CC signals represents a major step forward…

Google is quite agressive with Gemini CLI , offering 1000 free Gemini 2.5 Pro requests per day for individual users. Essentially trying to push Anthropic and OpenAI away.
https://blog.google/technology/developers/introducing-gemini-cli-open-source-ai-age…

Gemini CLI: your open-source AI agent
Free and open source, Gemini CLI brings Gemini directly into developers’ terminals — with unmatched access for individuals.

🤑 Introducing pay per crawl: enabling content owners to charge AI crawlers for access
#ai

Introducing pay per crawl: Enabling content owners to charge AI crawlers for access
Pay per crawl is a new feature to allow content creators to charge AI crawlers for access to their content.

Canada's "AI Minister" won't be far behind. (Or maybe he's already ahead... it _would_ explain a lot....) https://bsky.app/profile/did:plc:35jwgqplo2hl6v77qgsehmhe/post/3lvnyk3fo6s2c

The writing that Creative Commons is also "AI"-pilled has been on the wall for a while, and here we are are, wasting time and money on some useless signalling to "AI" scrapers that already don't care for any existing limitations…
https://creativecommons.org/2025/06/25/introducing-cc-signals-a-new-social-contract-for-the-age-of-ai/

Meta expands voice calls on WhatsApp to large businesses and explores AI-powered product recommendations for merchants' sites (Ivan Mehta/TechCrunch)
https://techcrunch.com/2025/07/01/meta-adds-busines…

Meta adds business voice calling to WhatsApp, explores AI-powered product reccomendations | TechCrunch
WhatsApp is adding more AI features to its business suite. The company on Tuesday announced it's introducing the ability for large businesses to reach

#Devstral, introducing the best #OpenSource model for #codingagents.

👀
https://www.theverge.com/news/695501/cloudflare-block-ai-crawlers-default

syslog-ng statement on #AI shortly after rsyslog AI announcement: https://fosstodon.org/@PCzanik/114856221034797034

🏢 Code Max: 5,000 messages/day for full-time development and multi-agent systems
⚡ Instant code generation with 131k-token context window and no weekly limits
https://www.cerebras.ai/blog/introducing-cerebras-code

Kontext: https://creativecommons.org/2025/06/25/introducing-cc-signals-a-new-social-contract-for-the-age-of-ai/

This https://arxiv.org/abs/2502.16644 has been replaced.
link: https://scholar.google.com/scholar?q=a

Mapping out AI Functions in Intelligent Disaster (Mis)Management and AI-Caused Disasters
This study maps the functions of artificial intelligence in disaster (mis)management. It begins with a classification of disasters in terms of their causal parameters, introducing hypothetical cases of independent or hybrid AI-caused disasters. We then overview the role of AI in disaster management and mismanagement, where the latter includes possible ethical repercussions of the use of AI in intelligent disaster management (IDM), as well as ways to prevent or mitigate these issues, which inclu…

Overly academic/distanced ethical discussions
Had a weird interaction with @/brainwane@social.coop just now. I misinterpreted one of their posts quoting someone else and I think the combination of that plus an interaction pattern where I'd assume their stance on something and respond critically to that ended up with me getting blocked. I don't have hard feelings exactly, and this post is only partly about this particular person, but I noticed something interesting by the end of the conversation that had been bothering me. They repeatedly criticized me for assuming what their position was, but never actually stated their position. They didn't say: "I'm bothered you assumed my position was X, it's actually Y." They just said "I'm bothered you assumed my position was X, please don't assume my position!" I get that it's annoying to have people respond to a straw man version of your argument, but when I in response asked some direct questions about what their position was, they gave some non-answers and then blocked me. It's entirely possible it's a coincidence, and they just happened to run out of patience on that iteration, but it makes me take their critique of my interactions a bit less seriously. I suspect that they just didn't want to hear what I was saying, while at the same time they wanted to feel as if they were someone who values public critique and open discussion of tricky issues (if anyone reading this post also followed our interaction and has a different opinion of my behavior, I'd be glad to hear it; it's possible In effectively being an asshole here and it would be useful to hear that if so).
In any case, the fact that at the end of the entire discussion, I'm realizing I still don't actually know their position on whether they think the AI use case in question is worthwhile feels odd. They praised the system on several occasions, albeit noting some drawbacks while doing so. They said that the system was possibly changing their anti-AI stance, but then got mad at me for assuming this meant that they thought this use-case was justified. Maybe they just haven't made up their mind yet but didn't want to say that?
Interestingly, in one of their own blog posts that got linked in the discussion, they discuss a different AI system, and despite listing a bunch of concrete harms, conclude that it's okay to use it. That's fine; I don't think *every* use of AI is wrong on balance, but what bothered me was that their post dismissed a number of real ethical issues by saying essentially "I haven't seen calls for a boycott over this issue, so it's not a reason to stop use." That's an extremely socially conformist version of ethics that doesn't sit well with me. The discussion also ended up linking this post: https://chelseatroy.com/2024/08/28/does-ai-benefit-the-world/ which bothered me in a related way. In it, Troy describes classroom teaching techniques for introducing and helping students explore the ethics of AI, and they seem mostly great. They avoid prescribing any particular correct stance, which is important when teaching given the power relationship, and they help students understand the limitations of their perspectives regarding global impacts, which is great. But the overall conclusion of the post is that "nobody is qualified to really judge global impacts, so we should focus on ways to improve outcomes instead of trying to judge them." This bothers me because we actually do have a responsibility to make decisive ethical judgments despite limitations of our perspectives. If we never commit to any ethical judgment against a technology because we think our perspective is too limited to know the true impacts (which I'll concede it invariably is) then we'll have to accept every technology without objection, limiting ourselves to trying to improve their impacts without opposing them. Given who currently controls most of the resources that go into exploration for new technologies, this stance is too permissive. Perhaps if our objection to a technology was absolute and instantly effective, I'd buy the argument that objecting without a deep global view of the long-term risks is dangerous. As things stand, I think that objecting to the development/use of certain technologies in certain contexts is necessary, and although there's a lot of uncertainly, I expect strongly enough that the overall outcomes of objection will be positive that I think it's a good thing to do.
The deeper point here I guess is that this kind of "things are too complicated, let's have a nuanced discussion where we don't come to any conclusions because we see a lot of unknowns along with definite harms" really bothers me.

Using Generative AI in Software Design Education: An Experience Report
Victoria Jackson, Susannah Liu, Andre van der Hoek
https://arxiv.org/abs/2506.21703 …

Using Generative AI in Software Design Education: An Experience Report
With the rapid adoption of Generative AI (GenAI) tools, software engineering educators have grappled with how best to incorporate them into the classroom. While some research discusses the use of GenAI in the context of learning to code, there is little research that explores the use of GenAI in the classroom for other areas of software development. This paper provides an experience report on introducing GenAI into an undergraduate software design class. Students were required to use GenAI (in …

1/2: Eeek!
https://blog.google/products/chrome/store-reviews/
Just me, or is this a horrifying prospect? The SEO wars and the Amazon homepage have taught us that businesses will do anything - ANYTHING, ethics be damned, to game their online reputations. This …

Chrome is helping you shop smarter and safer with store reviews.
We’re introducing store reviews to Google Chrome in the U.S. to make your online shopping experience safer and more efficient. This feature uses data from Google Shopping and other popular review websites to provide people with helpful insights when shopping on e-commerce sites.Just click the icon to the left of the web address and you can access an AI-generated summary of reviews. The description will cover topics like customer service, product quality, shipping, pricing and returns, helping…

Creative Commons jumps the shark.
*smdh*
#AI #CreativeCommons https://mastodon.social/@creativeco…

MAEBE: Multi-Agent Emergent Behavior Framework
Sinem Erisken (Independent Researcher), Timothy Gothard (Independent Researcher), Martin Leitgab (Independent Researcher), Ram Potham (Independent Researcher)
https://arxiv.org/abs/2506.03053

MAEBE: Multi-Agent Emergent Behavior Framework
Traditional AI safety evaluations on isolated LLMs are insufficient as multi-agent AI ensembles become prevalent, introducing novel emergent risks. This paper introduces the Multi-Agent Emergent Behavior Evaluation (MAEBE) framework to systematically assess such risks. Using MAEBE with the Greatest Good Benchmark (and a novel double-inversion question technique), we demonstrate that: (1) LLM moral preferences, particularly for Instrumental Harm, are surprisingly brittle and shift significantly …

Demystifying the 7-D Convolution Loop Nest for Data and Instruction Streaming in Reconfigurable AI Accelerators
Md Rownak Hossain Chowdhury, Mostafizur Rahman
https://arxiv.org/abs/2507.20420

Demystifying the 7-D Convolution Loop Nest for Data and Instruction Streaming in Reconfigurable AI Accelerators
Convolution remains the most compute-intensive operation in AI acceleration, often constituting over 80-90% of the workload. Existing approaches in spatial architectures such as coarse-grained reconfigurable arrays (CGRAs) and field-programmable gate arrays (FPGAs) frequently rely on loop unrolling or GEMM-based matrix transformations, introducing significant overhead in both data movement and instruction control. This paper presents a new framework designed to systematically demystify the 7-di…

This looks like great fun; big big construction machines driven by AI - that's what we want our AI to be doing, driving huge machines with shovels on.
That must be great fun to debug.
https://bedrockrobotics.com/news/introducing-bedrock-robotics…

Bedrock Robotics

Nice new models from Mistral
https://mistral.ai/news/voxtral?utm_source=superhuman&utm_medium=newsletter&utm_campaign=thinking-machines-lan…

Voxtral | Mistral AI
Introducing frontier open source speech understanding models.

Critically-Damped Higher-Order Langevin Dynamics
Benjamin Sterling, Chad Gueli, M\'onica F. Bugallo
https://arxiv.org/abs/2506.21741 https://

Critically-Damped Higher-Order Langevin Dynamics
Denoising Diffusion Probabilistic Models represent an entirely new class of generative AI methods that have yet to be fully explored. Critical damping has been successfully introduced in Critically-Damped Langevin Dynamics (CLD) and Critically-Damped Third-Order Langevin Dynamics (TOLD++), but has not yet been applied to dynamics of arbitrary order. The proposed line of work generalizes Higher-Order Langevin Dynamics (HOLD), a recent state-of-the-art diffusion method, by introducing the concept…

Introducing Horizon Alpha, a new stealth #LLM 🌅
currently #FREE 👀
https://openrouter.ai/openrouter/horiz

In the #ISE2025 lecture today we were introducing our students to the concept of distributional semantics as the foundation of modern large language models. Historically, Wittgenstein was one of the important figures in the Philosophy of Language stating thet "The meaning of a word is its use in the language."

[Thread] Cluely unveils a desktop AI assistant that it says can help users cheat on meetings, sales, lectures, interviews, learning new software, and more (Roy/@im_roy_lee)
https://x.com/im_roy_lee/status/1938718987975827651

Roy (@im_roy_lee) on X
introducing @cluely. today is the start of a world where you never have to think again. we just killed 9 industries (thread):

I'm not shocked that Musk would try to manipulate Grok to vent his opinions. We've seen his manipulation before in the twitter recommender system.
I am kind of fascinated by how difficult it is. I think that broadly trained LLMs tend to converge to the same "worldview", which is mostly left-libertarian.
You can't really force them away from this on one issue without introducing inconsistencies, and breaking the guardrails.

Agentic Satellite-Augmented Low-Altitude Economy and Terrestrial Networks: A Survey on Generative Approaches
Xiaozheng Gao, Yichen Wang, Bosen Liu, Xiao Zhou, Ruichen Zhang, Jiacheng Wang, Dusit Niyato, Dong In Kim, Abbas Jamalipour, Chau Yuen, Jianping An, Kai Yang
https://arxiv.org/abs/2507.14633…

Agentic Satellite-Augmented Low-Altitude Economy and Terrestrial Networks: A Survey on Generative Approaches
The development of satellite-augmented low-altitude economy and terrestrial networks (SLAETNs) demands intelligent and autonomous systems that can operate reliably across heterogeneous, dynamic, and mission-critical environments. To address these challenges, this survey focuses on enabling agentic artificial intelligence (AI), that is, artificial agents capable of perceiving, reasoning, and acting, through generative AI (GAI) and large language models (LLMs). We begin by introducing the archite…

Exploiting Jailbreaking Vulnerabilities in Generative AI to Bypass Ethical Safeguards for Facilitating Phishing Attacks
Rina Mishra, Gaurav Varshney
https://arxiv.org/abs/2507.12185

Exploiting Jailbreaking Vulnerabilities in Generative AI to Bypass Ethical Safeguards for Facilitating Phishing Attacks
The advent of advanced Generative AI (GenAI) models such as DeepSeek and ChatGPT has significantly reshaped the cybersecurity landscape, introducing both promising opportunities and critical risks. This study investigates how GenAI powered chatbot services can be exploited via jailbreaking techniques to bypass ethical safeguards, enabling the generation of phishing content, recommendation of hacking tools, and orchestration of phishing campaigns. In ethically controlled experiments, we used Cha…

Tensor Manipulation Unit (TMU): Reconfigurable, Near-Memory Tensor Manipulation for High-Throughput AI SoC
Weiyu Zhou, Zheng Wang, Chao Chen, Yike Li, Yongkui Yang, Zhuoyu Wu, Anupam Chattopadhyay
https://arxiv.org/abs/2506.14364

Tensor Manipulation Unit (TMU): Reconfigurable, Near-Memory Tensor Manipulation for High-Throughput AI SoC
While recent advances in AI SoC design have focused heavily on accelerating tensor computation, the equally critical task of tensor manipulation, centered on high,volume data movement with minimal computation, remains underexplored. This work addresses that gap by introducing the Tensor Manipulation Unit (TMU), a reconfigurable, near-memory hardware block designed to efficiently execute data-movement-intensive operators. TMU manipulates long datastreams in a memory-to-memory fashion using a RIS…

Google DeepMind unveils Aeneas, an AI model for contextualizing ancient Latin inscriptions, to help historians interpret and restore fragmentary texts (Google DeepMind)
https://deepmind.google/discover/blog/aeneas-transforms-how-historians-connect-the…

Aeneas transforms how historians connect the past
Introducing the first model for contextualizing ancient inscriptions, designed to help historians better interpret, attribute and restore fragmentary texts.

Unraveling the Potential of Diffusion Models in Small Molecule Generation
Peining Zhang, Daniel Baker, Minghu Song, Jinbo Bi
https://arxiv.org/abs/2507.08005

Unraveling the Potential of Diffusion Models in Small Molecule Generation
Generative AI presents chemists with novel ideas for drug design and facilitates the exploration of vast chemical spaces. Diffusion models (DMs), an emerging tool, have recently attracted great attention in drug R\&D. This paper comprehensively reviews the latest advancements and applications of DMs in molecular generation. It begins by introducing the theoretical principles of DMs. Subsequently, it categorizes various DM-based molecular generation methods according to their mathematical and ch…

IndoorWorld: Integrating Physical Task Solving and Social Simulation in A Heterogeneous Multi-Agent Environment
Dekun Wu, Frederik Brudy, Bang Liu, Yi Wang
https://arxiv.org/abs/2506.12331

IndoorWorld: Integrating Physical Task Solving and Social Simulation in A Heterogeneous Multi-Agent Environment
Virtual environments are essential to AI agent research. Existing environments for LLM agent research typically focus on either physical task solving or social simulation, with the former oversimplifying agent individuality and social dynamics, and the latter lacking physical grounding of social behaviors. We introduce IndoorWorld, a heterogeneous multi-agent environment that tightly integrates physical and social dynamics. By introducing novel challenges for LLM-driven agents in orchestrating …