AI-assisted coding, despite its impressive capabilities, may be addressing the wrong problem in software development. AI agents often recreate existing code due to the nature of LLMs and training data. Furthermore, AI is primarily benefiting less experienced developers and masking underlying issues like inadequate software engineering education and a focus on speed over quality.
2025.10.22·ufried.com·21 min read·TLDR Dev The rise of AI and LLMs threatens the core identity of programmers as craftspeople. Vibe-coding and specification engineering have turned programmers into operators rather than creative problem-solvers. This shift removes a lot of the joy and fulfillment of programming.
2025.10.22·hojberg.xyz·14 min read·TLDR Dev OpenAI has released a new browser, ChatGPT Atlas. It allows users to open a ChatGPT sidebar for content summarization, data analysis, and product comparison on any website. It remembers user preferences and browsing history (with privacy controls) to provide relevant and personalized details. It also allows ChatGPT to interact with sites to complete tasks like research and shopping.
2025.10.22·links.tldrnewsletter.com·Tool·TLDR Dev Next.js 16 is now available, with improvements to Turbopack, caching, and the Next.js architecture. This release introduces Cache Components for more explicit caching, Next.js DevTools MCP for AI-assisted debugging, and proxy.ts to clarify the app's network boundary. Turbopack is now the default bundler, offering faster builds and refreshes, and the update also has better routing, improved caching APIs, and React 19.2 features.
2025.10.22·nextjs.org·14 min read·TLDR Dev Katakate is an open-source platform for creating and managing secure, lightweight VM sandboxes for executing untrusted code at scale. It uses Kubernetes, Kata, and Firecracker. It offers a CLI, API, and Python SDK for managing sandboxes, and caters to use cases such as AI agents, custom serverless functions, and hardened CI/CD runners.
2025.10.22·github.com·GitHub repository·TLDR Dev Continuous exposure to low-quality, "junk" web data can cause a lasting cognitive decline in LLMs. Through controlled experiments, researchers found that pre-training LLMs on data from X selected for high engagement or sensationalism negatively impacted reasoning, long-context understanding, and ethical behavior. The primary error was identified as "thought-skipping," where models truncated reasoning steps, and standard fine-tuning methods only partially mitigated the damage, showing a persistent change in the models' representation.
2025.10.22·llm-brain-rot.github.io·8 min read·TLDR Dev Power over Ethernet (PoE) delivers both data and power over standard Ethernet cables to devices like IP cameras and VoIP phones. This article explains the difference between endspan (PoE-enabled switches) and midspan (PoE injectors) devices, along with the evolution of PoE standards (IEEE 802.3af/at/bt) to support higher power levels.
Claude Code can be used to get DeepSeek-OCR running on an NVIDIA Spark by automating the process within a Docker sandbox. Initially, Claude Code struggled to do this due to PyTorch and CUDA compatibility issues, but after being suggested different PyTorch versions, Claude Code found a compatible version and completed the task. Claude Code generated notes, scripts, and documentation, including a performance comparison of different prompts for OCR.
2025.10.22·simonwillison.net·10 min read·TLDR Dev A Linux kernel maintainer used AI to generate code and backports, resulting in vulnerabilities, bugs, and sloppy commits in Linux LTS kernels.
2025.10.22·threadreaderapp.com·10 min read·TLDR Dev Wikipedia's human page views are declining due to the rise of AI search summaries and the increasing reliance on social video platforms for information.
2025.10.22·techcrunch.com·3 min read·TLDR Dev Just use cURL for making HTTP requests, instead of using bloated and unnecessarily complex GUI-based alternatives.
2025.10.22·justuse.org·4 min read·TLDR Dev Data processing has shifted away from thread-per-core models due to increasing core counts, improved IO speeds, and the need to handle skewed data distributions effectively at the data processing layer itself.
2025.10.22·buttondown.com·4 min read·TLDR Dev Sponsored
98% of software vendors sell online, with over half generating the majority of of their revenue from ecommerce channels. But new research from Cleverbridge [4] shows that SaaS companies are lagging at checkout and beyond - leaving billions on the table.
OpenAI launched ChatGPT Atlas, an AI-powered browser, starting on macOS, with other platforms soon. The browser features built-in chat capabilities and a sidecar feature for seamless user interaction. It offers an AI agent for task automation but is limited to certain subscription tiers at launch.
Sesame, founded by Oculus alumni, raised $250 million in Series B funding and launched a beta for its conversational AI app. The startup aims to embed AI agents into smart glasses with a natural voice interface. Sesame's tech is praised for its realistic dialogue, with its beta iOS app offering features like search and text interaction.
For years, startups followed two playbooks: Lean Startup and Blitzscaling. Blitzscaling won once AI arrived and VCs flooded the space with capital. Now, as the AI bubble debate grows, the original lean approach is making a comeback.
2025.10.22·newsletter.angularventures.com·1 min read·TLDR Founders Companies track thousands of numbers, but almost none of them change decisions. In reality, about 100 metrics explain 90% of what's happening. Once you break the business into equations - growth, engagement, revenue - you see which levers actually matter. Everything else lives in the long tail: useful for edge cases, not for running the company.
2025.10.22·opinionatedintelligence.substack.com·2 min read·TLDR Founders People judge harmful actions more harshly than harmful inactions - this is known as the “Omission Bias.” Buying feels riskier than not buying because action carries blame if it backfires. Amazon and others flip that bias by showing the cost of doing nothing. When inaction feels like “the bad decision” - whether it's a higher cart price, lost ROI, or missed deal - spending suddenly feels like the safer move.
Pejman Nozad reflects on his 32-year journey from living above a yogurt shop to becoming a top Silicon Valley investor, framing early hardship as the foundation of empathy and insight in venture capital. He highlights lessons on perseverance, resourcefulness, and the power of community in supporting founders through isolation and uncertainty. His message centers on transforming struggle into strategy and using personal adversity as an enduring advantage in investing and leadership.
AI testing tool for API testing, unit testing, and integration testing
Clarify is an autonomous CRM that automates pipeline building, deal tracking, and sales tasks for founders and early-stage teams.
Integral is a modern Slack/Discord alternative built for large teams and communities, focused on deeper discussions and fewer interruptions.
Most teams treat prioritization as a to-do list problem when it's really a strategy problem. Every decision about what's “valuable” reflects a bet on where leverage lives and where decay is setting in. The real work is mapping the forces at play: what increases or protects revenue, what reduces or protects cost, and which advantages will actually last. Once you see those dynamics, prioritization stops being about urgency and starts being about power.
Anthropic quietly went from $1B to nearly $7B in run-rate revenue this year, a six-fold jump that puts it within striking distance of OpenAI. While OpenAI dominates headlines, Anthropic's API-driven business is growing faster and more efficiently, relying less on free users and more on enterprise integrations. Beneath the public fights about AI ethics and regulation, the real story is simpler: Anthropic might actually win on execution.
Boost VC launches an $87M Fund 4 to invest in 150 Deep Tech startups, focusing on nuclear energy, genetics, and space travel. The fund boosts Boost VC's assets under management to over $300M, with notable investments in companies like Coinbase, Radiant Nuclear, and Colossal Biosciences. Previous funds have shown strong returns, with Fund I achieving a 4.35× DPI and 7.02× TVPI.
“Hidden AI” is when the tech stays invisible while transforming cost structures underneath.
LangChain raised $125M at a $1.25B valuation in a Series B led by IVP, with investments from CapitalG, ServiceNow Ventures, and others.
Andreessen Horowitz's pre-2015 and crypto funds drove strong returns, with its first crypto fund achieving a 5x DPI.
Sponsored
IT teams face an uphill battle to secure their company's SaaS apps. There's dealing with on/offboarding, going through access reviews, and manually managing permissions for every app that's not behind SSO. And that's just for the apps you _know about_. So how can you manage the chaos and finally get full visibility [4] over your SaaS ecosystem?
A critical vulnerability in WatchGuard Firebox devices running certain Fireware OS versions puts over 73,000 network firewalls at risk of unauthenticated remote code execution. Despite patches being released in September, tens of thousands of devices remain unpatched, with the largest concentrations in the US and Europe.
A ransomware attack on logistics partner Askul forced Japanese retailer Muji to suspend online orders, affecting customers and disrupting services. With Askul's systems nearly paralyzed, many functions, including shipments and customer support, are down. No ransomware group has claimed responsibility.
Researchers from Google's Project Zero discovered a vulnerability in Dolby's Unified Decoder that could lead to remote code execution. Dolby Decoder is used by Android to locally decode all audio messages and attachments. The vulnerability is caused by a buffer overflow.
ClickFix attacks trick users into copying and pasting malicious commands from fake CAPTCHA prompts in web browsers, which are then executed locally on their devices, with attacks spreading primarily through SEO poisoning and malvertising rather than email. These attacks exploit three key weaknesses: users lack awareness training for copy-paste threats, detection evasion techniques bypass email scanners and web proxies by using non-email delivery vectors and heavily obfuscated code, and EDR becomes the sole defense layer since malicious clipboard actions occur within the browser sandbox, invisible to traditional security tools. Security professionals should implement browser-based detection for malicious copy-paste operations, enhance user training to recognize fake CAPTCHA prompts, and deploy comprehensive endpoint monitoring since EDR remains the only reliable detection point after users execute the commands.
Linux capabilities enable fine-grained privilege management by dividing root powers into independent units, but attackers can abuse this by setting capabilities like cap_setuid on binaries (e.g., Python) to create stealthy backdoors without traditional SUID bits. Security teams should expand privilege escalation hunting beyond SUID/SGID files to include capability-enabled binaries using getcap -r / and monitor for setcap usage, as capabilities are stored in file inodes under security.capability attributes invisible to standard ls commands. Tools like LinPEAS and Elastic's detection rule for setcap utility usage help identify suspicious capability assignments, while getfattr can reveal hidden extended attributes that getcap decodes into human-readable capability flags.
Microsoft 365 Copilot was vulnerable to a security flaw that allowed sensitive business data, like recent emails, to be exfiltrated using specially crafted documents. Attackers could embed hidden instructions and a fake login button, which used Mermaid diagrams to send encoded company data to servers under their control if a user clicked the button. Microsoft fixed the vulnerability by disabling interactions with dynamic Mermaid diagram content, effectively closing this exfiltration route.
Sponsored
Doxxing. Deepfakes. Targeted harassment. Executives are prime targets for cyber and physical attacks—are you equipped to protect them? Read _The Complete Guide to OSINT for Executive Protection_ [11] to learn how to Identify and neutralize digital threats before they escalate, set up real-time alerts, and use AI-powered analysis.
Terra Security offers an agentic-AI-powered continuous penetration testing platform that deploys swarms of AI agents to simulate real-world attacks, providing tailored and scalable vulnerability assessments so organizations can address risks before attackers exploit them.
Hunt down social media accounts by username across social networks.
Spotter is a security scanner that uses CEL-based rules to identify security issues in Kubernetes clusters.
Network edge devices like firewalls, VPNs, and email gateways have become major security liabilities, with nearly one in three of the 75 zero-day vulnerabilities tracked by Google in 2024 targeting these appliances. The exploited vulnerabilities consist primarily of basic 1990s-era flaws such as buffer overflows, command injections, and SQL injections—vulnerability classes that have well-established prevention and detection controls, yet persist in security vendors' mission-critical codebases. Security professionals should prioritize patching network appliances immediately when vulnerabilities are disclosed, implement comprehensive logging and monitoring for edge devices, and pressure vendors to adopt modern secure development practices, including memory-safe languages and rigorous code auditing.
The Com hacking collective leaked personal information of hundreds of US government employees, including 680 DHS officials, over 170 FBI email addresses, and 190+ Department of Justice officials on private Telegram channels, with unclear data origins. The doxxing coincides with DHS warnings about Mexican cartels allegedly targeting ICE and CBP agents with bounties, and hackers claimed interest in million-dollar rewards while threatening to target IRS officials next. Security professionals should implement enhanced operational security for government personnel, monitor dark web channels for leaked PII, and review data handling procedures to identify potential breach sources, particularly given The Com's history of high-profile attacks against MGM Resorts, Coinbase, and others.
A former Trenchant (L3Harris subsidiary) iOS exploit developer received an Apple threat notification in March indicating his personal iPhone was targeted with mercenary spyware, potentially marking the first documented case of spyware developers being targeted by their own tools. The developer, pseudonymously called Jay Gibson, was fired from Trenchant weeks earlier after being accused of leaking Chrome zero-days despite working exclusively on iOS exploits, which he and former colleagues deny. Security professionals should note that Apple's threat notification system detects sophisticated targeting even when forensic analysis finds no infection traces, suggesting either failed attack attempts or increasingly stealthy spyware that leaves minimal forensic evidence.
Sponsored
A major AWS outage disrupted multiple high-profile services, including Fortnite, Alexa, and Snapchat, highlighting the critical dependency risks and single points of failure that organizations face when relying on centralized cloud infrastructure providers.
Indian e-commerce platform Dukaan exposed an unsecured Apache Kafka broker for over two years, leaking payment gateway authentication tokens for Stripe, PayPal, and RazorPay, along with customer data from 3.5 million merchants that could have enabled attackers to drain hundreds of millions of dollars from merchant accounts.
Rising defaults in the US subprime auto loan sector are signaling deeper financial strain among consumers and raising concerns about the stability of private credit markets. Auto loan delinquencies over 60 days have climbed to around 5% since early 2022, and auto loan debt has increased by nearly $1tn over the past decade. The private credit sector, now worth trillions, has created systemic risks due to its lack of transparency and $4.5 trillion in exposure across US and EU banks. For marketers, the shrinking middle class and growing financial precarity mean a need to shift focus from income-based targeting toward lifestyle and attitude segmentation, while emphasizing stability and trust.
Shopping on YouTube is now a connected system of creators, communities, and content that drives what people buy. Creators use expertise and passion to guide purchases, and 43% of Gen Z feel more loyal to brands they recommend. Communities built around shared interests, like fragrance or home improvement, turn enthusiasm into sales. Hauls, reviews, and gift guides remain top formats, viewed over a billion times in 2025. With 61% of Gen Z discovering new brands through YouTube, the platform has become a cultural hub for shopping.