The Slop Report - June 25, 2026
Today: Deepfake scams are getting uglier, and Bitdefender now…; Failing Robot Cop Company Knightscope Now Publishing…; Show HN: RealTube – Watch YouTube with filters for…
Your daily digest of AI-generated content news from around the web. All signal, no slop.
1. Deepfake scams are getting uglier, and Bitdefender now has an app for the panic
Digital Trends - · Jun 24
Bitdefender has introduced RealCheck, a deepfake detection app for Android and iOS designed to combat increasingly sophisticated video scams, including those involving impersonation and financial fraud. The app analyzes both uploaded files and web-hosted videos from various social platforms to assess their authenticity and potential for deceit, filling a critical need as social media becomes a primary avenue for such scams. As consumers struggle to identify deepfakes, RealCheck aims to provide a valuable tool for verifying the integrity of video content before users take action based on potentially misleading information.
2. Failing Robot Cop Company Knightscope Now Publishing Bizarre AI Slop Fan Fiction About Its Robots Solving Absurd Crimes
Futurism - · Jun 24
Knightscope, a robot security company, has faced significant failures since its inception 13 years ago, including a drastic stock plunge of over 99% since going public in 2022 and its robots failing to assist in actual security situations. In an unexpected pivot, the company has begun publishing a series of bizarre AI-generated fan fiction titled “The Knightscope Chronicles,” depicting its robots as crime-fighting heroes in absurd scenarios, highlighting both the disconnection from reality in Knightscope’s branding efforts and the potentially dire need for a re-evaluation of its product effectiveness and market relevance.
3. Show HN: RealTube – Watch YouTube with filters for AI-generated content
Hacker News - · Jun 24
RealTube has launched a new feature that allows users to filter AI-generated videos on its platform. This development is significant as it aims to differentiate authentic user-generated content from AI-generated material, addressing growing concerns over misinformation and content authenticity in the digital landscape. By implementing such filters, RealTube enhances user experience and promotes transparency in video consumption.
4. ChatGPT and Gemini could be quietly affecting your voting decisions, analysis shows
Digital Trends - · Jun 25
An analysis by The Washington Post has revealed that major AI chatbots, including OpenAI’s ChatGPT and Google’s Gemini, may exhibit political bias in their responses to political queries, with ChatGPT providing left-leaning answers 80% of the time while Gemini maintained a more balanced approach. Although the study does not directly indicate that these biases influence voter behavior, it highlights the potential for chatbots to frame political issues for users, which could affect how they engage with news and make decisions. These findings raise questions about the neutrality of AI platforms, which are increasingly used to understand complex political topics.
5. The CPU’s growing role in agentic AI infrastructure
The Register - · Jun 25
The article discusses the increasing importance of CPUs in the infrastructure of agentic AI, highlighting how they serve as the control plane that orchestrates data flow and workload management across cloud and datacenter environments. Companies like Arm, Google Cloud, AWS, and Microsoft Azure are developing custom Arm-based processors to optimize performance, security, and efficiency for AI applications. This shift towards purpose-built silicon illustrates a broader architectural trend in AI infrastructure where CPUs are essential for maximizing the potential of accelerators and ensuring the overall functionality of large-scale systems.
6. What 2 Content Strategists Say AI Should Never Write For You [Webinar] via @sejournal, @hethr_campbell
Search Engine Journal - · Jun 25
John Graham and Gabriel Dillon, strategists at Contentful, emphasize that AI should not handle all writing tasks, as it tends to produce content that merely rehashes existing material rather than offering original insights. They argue that while AI is effective for tasks like research and organizing ideas, human input is crucial for creating distinctive content that engages readers. This approach matters because it highlights the need for a balanced workflow that incorporates both AI efficiency and human creativity, ensuring content stands out in an oversaturated digital landscape.
7. OpenAI’s free GPT-5.5 model makes ChatGPT better at understanding context
Engadget - · Jun 25
OpenAI has released an update to its GPT-5.5 Instant model, enhancing ChatGPT’s ability to understand context and adapt to user queries. This upgrade, which became the default model in May, improves the model’s performance by reducing errors and allowing for better responses to complex questions, thereby enhancing user interaction. The advancements are significant as they aim to provide more relevant and refined information, improving the overall user experience with the chatbot.
8. Stripe, Anthropic, and OpenAI Are Backing Effort To Stop Respiratory Infections
Slashdot - · Jun 25
Stripe, alongside Anthropic, OpenAI, and other prominent funders, is launching a $500 million nonprofit named Intercept aimed at combating respiratory infections, including the common cold and flu. The initiative seeks to develop innovative prevention strategies such as vaccines and advanced air-cleaning technologies in public spaces. This effort is significant as it addresses the substantial societal burden of respiratory viruses, which have been historically neglected in favor of more lucrative pharmaceutical projects.
9. Calling everything AI-generated is lazy
Hacker News - · Jun 25
Frank Denis critiques the growing tendency to label various written works as “AI-generated” without substantive criticism, arguing that such comments are dismissive and unhelpful. He emphasizes the challenges faced by personal writers, especially non-native speakers, who use AI tools to improve their writing, and he calls for more constructive feedback rather than vague accusations, which can discourage authentic expression and risk creating a negative environment for content creation.
10. Google’s Gemini 3.5 Pro Release Slips to July
Hacker News - · Jun 25
Google has delayed the launch of its Gemini 3.5 Pro AI model to July, pushing it back from an initial June release date. This decision comes as the company seeks to incorporate feedback from early testers amid increasing competition from rivals like Anthropic and OpenAI, particularly in the area of coding capabilities. The updated model is expected to excel in long-horizon tasks and improve upon aspects criticized in the earlier Flash 3.5 model.
11. Cerebras stock plunges after earnings as CEO says margin outlook was misunderstood
TechCrunch AI - · Jun 24
Cerebras Systems experienced a nearly 20% drop in its stock price following its first earnings report as a public company, despite reporting a 94% year-over-year revenue increase to $193 million. The decline stemmed from a revised outlook on gross margins, which CEO Andrew Feldman attributed to investors misinterpreting the company’s need to rent back equipment from a major customer during its data center build-out, ultimately affecting profit margins for the year. This situation is significant as it reflects investor concerns regarding profitability amidst expansion efforts in the competitive AI chip market.
12. Gemini in Chrome can now see exactly what you’re looking at on screen
Digital Trends - · Jun 24
Google has introduced a new feature for its Gemini AI in Chrome called “Select from screen,” which allows users to highlight text or images on a webpage and send them directly to Gemini for more contextual interactions. This update aims to enhance Gemini’s functionality as an intelligent assistant rather than just a chatbot, enabling it to understand user intent and assist with tasks more effectively. The rollout coincides with a broader update, allowing developers to access computer use capabilities directly through Gemini 3.5 Flash, improving its utility in enterprise workflows and multi-step automation processes.
13. OpenAI and Broadcom announce chip designed for LLM inference at scale
Ars Technica - · Jun 24
OpenAI and Broadcom have unveiled a new chip, named Jalapeño, specifically engineered for large language model (LLM) inference in data centers, marking a significant advancement in AI infrastructure. Developed with insights from OpenAI’s roadmap, this ASIC promises improved performance per watt compared to current options, as both companies aim to enhance their capabilities amidst a global compute crunch. The deployment of Jalapeño chips is expected by the end of this year, reflecting a growing trend toward custom silicon solutions in the AI industry.
14. Meta Pauses Employee-Tracking Program Following Internal Data Leak
Slashdot - · Jun 24
Meta has halted its Model Compatibility Initiative (MCI), a program designed to track employee computer activity to train AI systems, after a data leak exposed sensitive information to more employees than intended. Despite the company’s claims of safeguarding privacy, concerns raised by employees about the program’s implications for personal data security led to significant internal backlash, prompting the pause until adequate protections can be assured. This incident highlights ongoing tensions within Meta regarding employee privacy and the balance between leveraging data for AI training and respecting individual rights.
15. AI Data Center Power Requirements in 2026: The Complete Grid-to-Chip Guide
Hacker News - · Jun 24
AI data centers are projected to demand between 100 to 750 megawatts (MW) of power per site by 2026, driven by the shift from training to continuous inference workloads. As global electricity demand for data centers is expected to surpass 1,000 terawatt hours, infrastructure limitations present significant challenges, impacting hyperscalers, enterprises deploying private AI clusters, and governments investing in AI infrastructure. This shift necessitates strategic planning for power supply, combining grid access with innovative solutions like onsite generation and advanced cooling systems.
16. AI researchers continue to leave Google for its rivals
TechCrunch AI - · Jun 24
Top AI researchers, including Jonas Adler and Alexander Pritzel from Google, are leaving the company to join Anthropic, continuing a trend of talent migration away from Google. This follows the departure of prominent figures like Noam Shazeer and John Jumper, raising concerns for Google as rivals like OpenAI and Anthropic ramp up their recruitment efforts ahead of potential public offerings. The exodus highlights a growing competition for AI expertise in the industry.
17. OpenAI and Broadcom Unveil Custom A.I. Chip Design
NY Times Tech - · Jun 24
The maker of ChatGPT, OpenAI, plans to utilize substantial energy resources, specifically consuming 10 gigawatts of electricity through its chip usage. This significant power demand highlights the vital role of energy consumption in AI development and raises concerns about the sustainability and environmental impact of large-scale AI operations.
18. OpenAI just made GPT-5.5 Instant more fun to talk to, and users may actually notice
Digital Trends - · Jun 24
OpenAI has updated its popular AI model, GPT-5.5 Instant, focusing on enhancing conversational quality and user engagement rather than just competing on technical benchmarks. The upgrade allows the AI to better understand user intent, making interactions feel more natural and tailored, which is important for tasks like planning and shopping. This shift may redefine how AI is perceived, emphasizing the importance of user experience in everyday applications.
19. Anthropic accuses Alibaba of running the largest distillation campaign yet against Claude
The Next Web - · Jun 24
Anthropic has accused Alibaba’s Qwen lab of conducting the largest known distillation campaign against its AI model, Claude, by utilizing nearly 25,000 fake accounts to engage in about 29 million exchanges that target valuable capabilities. This accusation comes amidst growing concern over foreign interference in the AI sector, with Anthropic’s letter to U.S. senators and White House officials highlighting the potential for Chinese firms to replicate advanced AI models without proper safety measures. The news has negatively impacted Alibaba’s stock and adds complexity to its ongoing legal battles related to national security designations.
20. Jalapeño is the first AI chip from OpenAI and Broadcom
Engadget - · Jun 24
OpenAI and Broadcom have jointly unveiled Jalapeño, their first custom AI chip designed specifically for large language model (LLM) inference. This collaboration aims to improve efficiency and performance in AI processing, with Jalapeño reportedly offering better performance per watt compared to current technologies. As OpenAI ventures into chip production, the launch signifies a strategic move to enhance the capabilities and accessibility of its AI tools, with initial deployments expected in late 2026.
21. ‘We should be worried’: report sheds light on ICE’s booming arsenal of hi-tech surveillance tools
The Guardian Tech - · Jun 24
A new report from the immigration rights organization Mijente and other researchers highlights a significant increase in U.S. government contracts for AI-powered surveillance technologies by ICE and CBP, doubling from $310 million in 2025 to a record $513 million in 2026. This surge in spending, primarily directed towards companies like Palantir and Anduril, reflects a concerted effort to enhance surveillance capabilities at the border and among immigrant populations, raising concerns over privacy and the increasing militarization of immigrant enforcement efforts. This growth in technical resources positions ICE as the most well-funded law enforcement agency in the U.S. and underscores the broader implications for civil rights and immigration policy.
22. Elon Musk loses trillionaire status as SpaceX and Tesla stock drops
The Guardian Tech - · Jun 24
Elon Musk has lost his status as the world’s first trillionaire due to a significant drop in stock prices for his companies, SpaceX and Tesla, following a broader market selloff triggered by concerns over interest rate hikes and an AI bubble. As of the market close, Musk’s net worth fell to approximately $970.2 billion, though he remains the wealthiest person globally, with a considerable lead over other billionaires. This decline underscores the volatility of wealth tied to stock prices, particularly in the tech sector, and highlights how quickly market dynamics can affect fortunes at this scale.
23. OpenAI Unveils First Chip As Part of Broadcom Deal
Slashdot - · Jun 24
OpenAI has introduced its first custom AI chip, named Jalapeno, in collaboration with Broadcom, aimed at optimizing inference for ChatGPT and other AI applications. This custom chip, described as an “Intelligence Processor,” is part of OpenAI’s strategy to develop a comprehensive tech stack to enhance efficiency and accessibility in AI services, with initial deployment expected by the end of
- This launch marks a significant shift towards proprietary hardware in AI, potentially reducing reliance on current suppliers like Nvidia.
24. Microsoft’s new Surface PCs are cheaper — but there’s a catch
Digital Trends - · Jun 24
Microsoft has unveiled new entry-level models of its Surface Pro and Surface Laptop, offering lower starting prices by reducing RAM to 8GB while retaining existing processors and storage. This change, aimed at making the devices more affordable, means that these new models will not qualify for Microsoft’s Copilot+ certification, which requires at least 16GB of RAM, thus limiting access to advanced AI features. The move highlights the ongoing challenges in the tech industry regarding rising component costs and the implications for future software capabilities.
25. Walmart’s First Nuclear Deal Shows Demand Beyond AI Data Centers
Slashdot - · Jun 24
Walmart has signed a long-term contract with Constellation Energy to purchase nuclear power from an Illinois plant, marking the retail giant’s first foray into nuclear energy. The agreement, which amounts to 176 megawatts over 15 years, supports Walmart’s commitment to achieving net carbon emissions elimination from its U.S. operations by 2040 and showcases a growing demand for sustainable energy sources beyond just powering AI data centers. This deal will also facilitate an expansion of the plant’s capacity, underscoring the broader potential for nuclear energy in corporate sustainability efforts.
25 stories sourced from Ars Technica, Digital Trends, Engadget, Futurism, Hacker News, NY Times Tech, Search Engine Journal, Slashdot, TechCrunch AI, The Guardian Tech, The Next Web, The Register. The Slop Report is published daily. Subscribe via RSS.