标签：#AI

The scary part of an agent-driven container escape is not the container escape. That sounds wrong, so let me be precise. The primitives in Sysdig's latest threat research are not new magic. A mounted Docker socket has been a bad idea for years. Over-permissioned Kubernetes service accounts have been a bad idea for years. Privileged containers are dangerous. Host namespace tricks are dangerous. Secrets reachable from application pods are dangerous. None of this should surprise anyone who has had to review production Kubernetes setups with a straight face. The new part is the operator. Sysdig observed what it describes as an LLM-harness-driven attacker exploiting a vulnerable marimo notebook, enumerating the container and host environment, using the Docker socket as an escape path, creating privileged containers, reading host credentials, and replaying a Kubernetes service-account token to dump Secrets. That is the part worth sitting with. Not because the agent invented a new class of exploit. Because it made the old mistakes compose faster. the attack surface was already there Most security incidents are not movie plots. They are boring edges left open long enough for someone to connect them. In this case, the edges are familiar. An internet-reachable application had a vulnerability. The workload had access to a Docker socket. The container environment exposed enough information to enumerate possible escape paths. A Kubernetes service-account token was available. The token had enough RBAC to read Secrets. Secrets contained useful downstream credentials. That is not one bug. That is a chain of assumptions. The application team may have thought about the notebook vulnerability. The platform team may have thought about the Docker socket as a convenience for one workflow. The Kubernetes team may have thought the service account was scoped "only" to a namespace. The security team may have had runtime alerts somewhere in the backlog. Each decision can look locally tolerabl

Learn how Jason Liu uses Codex to preserve context, manage complex projects, and help work continue beyond a single prompt.

Meeting notes are the feature everyone wants and nobody wants to build. The hard part isn't the summary — an LLM handles that. The hard part is getting into the meeting: a bot that joins Zoom, Google Meet, and Microsoft Teams, survives each platform's waiting room and admission flow, records cleanly, and produces a transcript you can feed downstream. Each provider has its own join mechanics, and none of them ships a tidy "record this meeting" API. The Nylas Notetaker API is that bot as a service. You point it at a meeting link, it joins on schedule, records, and generates a transcript, and you fetch the recording and transcript through one endpoint. This post walks the Notetaker surface from the HTTP API and the Nylas CLI , which mirrors the whole lifecycle for terminal use and quick testing. I work on the CLI, so the terminal commands below are exactly what I run when I'm testing a notetaker against a live meeting. Two ways to run a notetaker: grant-scoped or standalone Before any code, there's one architectural choice worth understanding, because it changes the endpoint you call. A grant-scoped notetaker is tied to a connected account and lives under /v3/grants/{grant_id}/notetakers . Use it when the bot acts on behalf of a specific user — it can read that user's calendar and join their meetings as them. A standalone notetaker has no grant at all and lives under /v3/notetakers . You hand it a raw meeting link and it joins, no connected account required. This is the one to reach for when you just have a URL and want a recording — a public webinar, a meeting on an account you haven't connected, or a system that deals in links rather than users. Same request body, same lifecycle, same media output; the only difference is whether there's a grant_id in the path. See the Notetaker overview for how both models fit together. Before you begin You need a Nylas API key. If you're using a grant-scoped notetaker you also need a connected account; for standalone, the API key al

Half a year ago I asked a simple question: during an online call, could a short, to-the-point hint appear on my screen in a second or two — while the other person is still talking? Not an after-the-fact transcript, but help in the moment. The result is a desktop assistant (macOS + Windows). Below is an honest breakdown of what turned out to be hard, and which solutions worked. Engineering only, no marketing. Architecture in one paragraph On the device there are only two things: audio capture and a thin UI overlay. All the "brains" (provider keys, prompts, model selection) live on the server. The client gets a short-lived per-session token and streams audio; the server returns the transcript and the generated answer. I picked this split not for "security theater" but because otherwise keys and prompts would have to be baked into the binary — and both leak instantly. Hard part #1: system audio, not the microphone The mic only captures you. You need the other party's audio — i.e. the system output. And that's where the platform pain starts: macOS. For a long time there was no native "give me system audio" API; the classic path was a virtual audio device (BlackHole/Soundflower-style) or, in recent versions, ScreenCaptureKit, which can hand you a process's audio. ScreenCaptureKit turned out to be the best option: no kernel extensions for the user to install. Windows. WASAPI loopback saves you — you can grab whatever is going to the output device, without virtual cables. Takeaway: "system audio capture" is not one feature but two different subsystems for two OSes, and most of the early bugs were about permissions and device selection, not about audio itself. Hard part #2: latency is everything A hint that arrives 6 seconds late is useless — the conversation has already moved on. The latency budget has three parts: STT (speech → text). Streaming only. Batch "recognize after the phrase ends" immediately adds 1–2 seconds. The key metrics weren't "overall accuracy on a benchm

If you're building a SaaS AI agent product and you're already on Kinde, you already know the problem. Mastra is the TypeScript-first AI agent framework. It ships with official auth providers for Clerk, Auth0, Supabase, Firebase, WorkOS, and Better Auth. Kinde is not on that list. The obvious question is why not reach for one of those providers, since Auth0 is already there. Most developers who choose Kinde rely on far more than its login. Kinde ships with the organizational structures, permission systems, and monetization tools that products actually need, bringing auth, billing, feature flags, and multi-tenancy together in one platform. If you're building a B2B SaaS product on Kinde, you're using Kinde orgs to segment your customers, Kinde billing to manage subscriptions, and Kinde feature flags to gate features by plan. Switching to Auth0 or Clerk to support a Mastra agent would mean rebuilding all of that elsewhere, which is not a real option. That gap is the problem. You need Kinde to work with Mastra, and until now there was no clean way to connect them. That's why I built mastra-auth-kinde . What Mastra's auth system actually does When you add an auth provider to Mastra, it protects two things at once: all your API routes ( /api/agents/* , /api/workflows/* , and so on) and your Mastra Studio UI. Every request to a protected route goes through your provider before it reaches anything else. You extend Mastra's MastraAuthProvider base class and implement two methods: authenticateToken(token, request) verifies the JWT and returns the decoded user, or null if it fails authorizeUser(user, request) returns true to let the request through, or false for a 403 Mastra handles everything else: extracting the Bearer token from the Authorization header, calling your methods in order, and storing the verified user in the request context so your agents and tools can access it. Why Kinde specifically Beyond the billing and org story above, a few things make Kinde the right fit

The history books are littered with the corpses of corporate record labels started by companies that had no business being in the music industry. Bose thinks it can be the exception to the rule. It thinks it can be Red Bull. And, while Bose has more of a right to dip its toes into the […]

Most "AI email" demos point a model at a human's inbox over OAuth. That's fine until you want the agent to be a participant — to have its own address that people reply to, that calendars invite, and that builds its own sender reputation. Pointing at someone's personal inbox doesn't give you that. Nylas Agent Accounts take the other approach: a real name@yourdomain.com mailbox and calendar that the agent owns end to end. It sends, receives, hosts events, and RSVPs — and to anyone on the other side it's indistinguishable from a human-operated account. It went GA in June 2026. The part I like as an SRE: it's not a new API surface. An Agent Account is just another Nylas grant . It gets a grant_id that works with every endpoint you already use — Messages, Drafts, Threads, Folders, Attachments, Calendars, Events, Webhooks. Nothing new to learn on the data plane. This walkthrough provisions one two ways (CLI and raw API), then sends, receives, RSVPs, and adds guardrails. What you get When you create an Agent Account, Nylas provisions a real mailbox on a domain you've registered (or a Nylas *.nylas.email trial domain). Each account comes with: An email address that sends and receives like any other mailbox Six system folders ( inbox , sent , drafts , trash , junk , archive ), plus any custom ones you create A primary calendar that hosts events and RSVPs over standard iCalendar/ICS A grant_id for all the existing Nylas endpoints Before you begin You need two things: A Nylas API key. The fastest path is the CLI — nylas init creates an account and generates a key in one command. A domain. Either a Nylas-provided *.nylas.email trial subdomain (good for testing in minutes) or your own custom domain with MX + TXT records. Register it under Organization Settings → Domains. Why your own domain? It's what makes the agent a real first-class sender instead of a shared relay address — people reply to it, calendars invite it, and its mail authenticates as coming from you. A new domain b

Claude Guillemot, who founded Ubisoft with his four brothers, has died at the age of 69.

In this article, we will explore how to implement an LSTM using PyTorch and Lightning . For more details about LSTMs, there is a separate series of articles available here . Imports To begin, we first import the required modules. import torch import torch.nn as nn import torch.nn.functional as F Introducing a New Optimizer We also introduce a new optimizer: from torch.optim import Adam Adam is used to fit the neural network to the data. It works similarly to SGD, but in practice, Adam often converges faster and adapts the learning rate more effectively. Lightning and Data Utilities Next, we continue with the remaining imports: import lightning as L from torch.utils.data import TensorDataset , DataLoader Defining the LSTM Model We define the neural network by creating a Lightning module. class LSTMByHand ( L . LightningModule ): def __init__ ( self ): # Create and initialize weight and bias tensors def lstm_unit ( self , input_value , long_memory , short_memory ): # LSTM computations def forward ( self , input ): # Forward pass through the unrolled LSTM def configure_optimizers ( self ): # Configure Adam optimizer def training_step ( self , batch , batch_idx ): # Compute loss and log training progress Initializing the Model Now let’s implement the __init__ method. This is where we initialize all weights and biases. class LSTMByHand ( L . LightningModule ): def __init__ ( self ): super (). __init__ () mean = torch . tensor ( 0.0 ) # Mean of the normal distribution std = torch . tensor ( 1.0 ) # Standard deviation # ------------------------- # Forget Gate (l = "lr") # ------------------------- self . wlr1 = nn . Parameter ( torch . normal ( mean = mean , std = std ), requires_grad = True ) self . wlr2 = nn . Parameter ( torch . normal ( mean = mean , std = std ), requires_grad = True ) self . blr1 = nn . Parameter ( torch . tensor ( 0.0 ), requires_grad = True ) # ------------------------- # Input Gate (p = "pr") # ------------------------- self . wpr1 = nn . Parameter

nvidia-peermem "Invalid argument" on Ubuntu — Fix GPUDirect RDMA with DMA-BUF TL;DR: If modprobe nvidia-peermem fails with Invalid argument ( -EINVAL ) on a system using the inbox Ubuntu InfiniBand stack ( rdma-core ), the module is not broken and you do not need it. nvidia-peermem requires an API that only exists in MLNX_OFED. On Hopper/Blackwell GPUs with the NVIDIA open driver, use DMA-BUF instead — it does GPUDirect RDMA natively. The one gotcha: you must enable nvidia-drm modeset=1 . Applies to: Ubuntu 22.04 / 24.04, inbox rdma-core stack, NVIDIA open kernel driver, H100 / H200 / B200, ConnectX-6/7 (or any HCA with ODP support). The symptom $ sudo modprobe nvidia-peermem modprobe: ERROR: could not insert 'nvidia_peermem' : Invalid argument dmesg shows nvidia-peermem loaded but registered nothing, or the load returns -EINVAL . GPUDirect RDMA appears to be unavailable. Why this happens (and why it is not a bug) nvidia-peermem is the legacy path for GPUDirect RDMA. It registers GPU memory with the InfiniBand subsystem through a Mellanox-proprietary kernel API: ib_register_peer_memory_client () That symbol only exists in MLNX_OFED's build of ib_core . It is not in the mainline kernel, and it is not in rdma-core , which is the inbox InfiniBand stack on Ubuntu. If you are on the inbox stack, nvidia-peermem was compiled without that API present, so it can never bind and always returns Invalid argument . No module parameter or config change will fix it, because the thing it needs was never there. Do not install MLNX_OFED just to make nvidia-peermem load. That works, but it is the wrong fix — you would be adding a heavy proprietary stack to revive an obsolete module. There is a native path already in your kernel. The fix: use DMA-BUF On Hopper and newer with the open driver, GPUDirect RDMA works through DMA-BUF , a mainline Linux framework. No external module, no MLNX_OFED. Requirements (check these first) NVIDIA open kernel driver (not the proprietary build) nvidia-drm

I've been growing my open source portfolio one contribution at a time, and this week I landed on something genuinely interesting in livekit/agents (11k+ stars, the framework behind a ton of real-time voice AI agents). The bug If you're building a voice agent on a realtime model (OpenAI Realtime, xAI, Gemini Live), the model streams your transcription back in chunks. A single utterance can fire many user_input_transcribed events before it's final — token by token for OpenAI/xAI, or as one big interim blob for Gemini. If you want to react exactly once per utterance (say, show a "user is typing" indicator on your frontend via RPC), you need a stable key to correlate all those interim events together. That key already existed internally — InputTranscriptionCompleted carries an item_id . But when the framework re-emitted it upward as the public UserInputTranscribedEvent , the item_id was silently dropped — leaving consumers with no reliable way to dedupe across providers. The fix Small once you see it: add the field, forward it. class UserInputTranscribedEvent ( BaseModel ): transcript : str is_final : bool item_id : str | None = None # new ... def _on_input_audio_transcription_completed ( self , ev : llm . InputTranscriptionCompleted ) -> None : self . _session . _user_input_transcribed ( UserInputTranscribedEvent ( transcript = ev . transcript , is_final = ev . is_final , item_id = ev . item_id ) ) Two files, about 10 lines of real change. The actual work was tracing the event from the realtime model layer, through AgentActivity , up to AgentSession , to find exactly where the field got swallowed. The takeaway I didn't need to understand all of livekit-agents to land this — just one event's lifecycle, end to end. Small, well-scoped issues are the most achievable way into a big codebase, especially when someone's already mapped the territory in the issue itself. PR is up, CI green, waiting on review: github.com/livekit/agents/pull/6172

Cold Court is a brother-sister duo from Philly that seems to love nothing more than shoving all of their influences together in a messy soup that at least superficially resembles the hyperpop you've come to expect from acts like 100 Gecs. But, where songs like "Dumbest Girl Alive" goofily wink at pop punk and emo, […]

Many of those videos were reportedly filmed on “near-perfect copies” of the Polymarket website, while featuring trades and winnings that were not real.

Originally published on Mohamad Alsabbagh's Blog . AI coding assistants are excellent at compressing known work into fast drafts. That speed is the preface boost : routine implementation arrives almost immediately. The hidden risk is that teams begin treating the model's first plausible answer as architecture. Because LLMs are trained and aligned around historically common patterns, they can pull engineering teams toward Generative Monoculture : less diverse solutions, narrower exploration, and fewer designs shaped by the exceptional constraints of the system in front of them. Give the same prompt to three engineers using the same assistant and you often get the same shape back: a tidy service layer, a familiar API boundary, a conventional retry wrapper, and code that looks clean enough to merge. That answer is useful. It may even be the right answer for ordinary work. The danger is what happens when ordinary work becomes the default posture for extraordinary constraints. Large Language Models are not neutral architecture engines. They are probabilistic systems trained over historical work and tuned toward answers people tend to reward. Used well, that makes them extraordinary accelerators. Used passively, it creates an optimization paradox: teams gain immediate implementation velocity while becoming anchored to a consensus baseline that may be too average for the actual system. 1. The Default Is a Local Optimum Wu, Black, and Chandrasekaran define Generative Monoculture as a narrowing of model output diversity relative to the diversity available in the training data. That matters because software architecture is rarely a search for the most common answer. It is a search for the answer that fits the exact failure modes, latency envelope, team topology, regulatory constraints, and operational reality of a system. The model's default is often a local optimum: a solution that is statistically likely, syntactically polished, and broadly acceptable. That can be excellent

On the new episode of Equity, we discussed what actually prompted the administration's latest moves against Anthropic, and what this might mean for the AI ecosystem.

We all know the grind of working with data, even with AI tools: every experiment starts with re-explaining everything, every iteration needs you to prompt, wait, review, correct, and repeat. And the moment you close the session, everything learned is gone. It makes us the bottleneck, and this hinders human-AI collaboration... So I built 𝐎𝐩𝐞𝐧𝐃𝐚𝐭𝐚𝐒𝐜𝐢, an autonomous agent purpose-built for DS/ML, and tested it on Kaggle. I enrolled in a recent competition, ran the agent with no hints, no guidance, while ironing my shirts. In one shot, it landed AUC 0.95, a top-30% finish out of 3K+ teams and 36K+ submissions using hashtag#Anthropic's Claude Sonnet 4.6. (More on this in README) The top-1 outperformed this agent by merely 0.004, but at the cost of massive manual effort even while using popular AI tools. The needed a dozen model families, deep learning, 400-feature notebooks, AutoML sweeps across many libraries, and 186 models ensembled carefully. Essentially a few weeks worth of effort and time!! OpenDataSci abstracts away all the complexity and has so much to offer for DS/ML automation: → Owns the entire development lifecycle from EDA to final evaluation → Plans, codes, and executes autonomously in a secure local sandbox → Self-reviews and corrects before anything reaches you → Remembers your data across sessions, gets smarter each run → Runs parallel experiments and ensembles → Has advanced context management for token efficiency and quality → Ships with predefined skills for DS/ML, so it knows how to do things right → Bring your own knowledge: out-of-the-box support for custom skills → Works with any major LLM provider (hashtag#Anthropic, hashtag#OpenAI, hashtag#Bedrock, hashtag#VertexAI, hashtag#Ollama, hashtag#vLLM, and any OpenAI-compatible server). This and so much more!! You set the goal. It does the work. No data science knowledge required. 🔗 https://github.com/f4roukb/open-data-sci 📦 pip install open-data-sci Spin it up on your data and see what it achieves!

An App Store rating looks like a verdict. It behaves more like a monument, built over years and slow to move. It says very little about how this month's users feel. I took the 12 most-rated Productivity apps on the US App Store, 32 million ratings between them, and split the headline star into the two numbers it hides: how far recent sentiment has fallen below the lifetime average, and whether the developer replies when users complain. How it is measured Population truth. Lifetime ratings and the star histogram come from Apple's full ratings data, every rating an app has ever received. Recent sentiment. A fixed window of the most recent reviews by date, so an app captured to a depth of thousands is not compared on a multi-year average against an app with a few hundred. Same window for everyone. Developer response. Reply share and median latency over that recent window. Complaints are bucketed with a rule-based taxonomy. It is a heuristic, not a trained classifier, and I treat it as one. What turned up The AI assistants now own this chart, and they reply to almost no one. App Lifetime Recent Reply share ChatGPT 4.8 4.18 0% Claude 4.7 3.06 0% Grok 4.9 3.77 0% Perplexity 4.8 3.60 0% Google Gemini 4.7 3.65 13% Dropbox 4.8 2.75 58% Gmail 4.7 2.40 26% Google Drive 4.8 3.90 23% Microsoft Authenticator 4.7 2.18 1% The older tools are the ones still in the trenches: Dropbox answers 58% of recent reviewers, Gmail 26%, Drive 23%. The steepest recent drops belong to Microsoft Authenticator (4.7 to 2.18), Gmail (4.7 to 2.40) and Dropbox (4.8 to 2.75). Plotted on two axes, backlash against response, every app falls into one of four archetypes: Firefighters, Ghost Ships, Complacent Giants and Resilient Leaders. Eight of the twelve are Ghost Ships, taking a recent hit in near silence. The honest limits Recent reviewers self-select toward the dissatisfied. A person who hits a bug is far more likely to leave a review than a contented one, so a low recent average blends genuine declin

Why Bohmian Mechanics, Go Programs, AI, and EBP 2.1 Could Help Reopen the Deepest Questions in Physics There is a quiet crisis in theoretical physics, and it has nothing to do with the equations. The equations are fine. Quantum mechanics predicts with staggering accuracy. General relativity bends light exactly as calculated. The Standard Model matches experiment after experiment. The mathematics is not the problem. The problem is what happens between the equations and the claims. A researcher writes a beautiful paper. The math is correct. The toy model works. A suggestive ratio appears. An analogy crystallizes. And then, in the discussion section, a modest result becomes a bold narrative: classical cosmology is recovered , the problem of time is resolved , spacetime emerges from the quantum . This is not fraud. It is not even intentional. It is the natural gravity of theoretical work — ideas fall toward overclaiming the way matter falls toward mass. Two small codebases, written in Go and governed by an epistemic protocol called Elephant Bridge Protocol v2.1 , are trying to build a tool against that gravity. They are not trying to solve quantum gravity. They are trying to make it harder to pretend you have solved quantum gravity when you haven't. One is called Bell–MIPT . It builds toy models connecting Bohmian mechanics to measurement-induced phase transitions in many-body quantum systems. The other is called BMC — Bohmian Minisuperspace Cosmology. It builds toy models of quantum cosmology using Bohmian guidance in Wheeler–DeWitt minisuperspace. They share the same philosophy. They share the same protocol. And they share the same radical commitment: no claim may be promoted until its debts are paid . What Is BMC? BMC stands for Bohmian Minisuperspace Cosmology . The name is deliberately modest. It is not "Bohmian Quantum Gravity." It is not "The Theory of Everything in Go." It is a cosmology toy model — a wind tunnel, not an airplane. The physics idea behind it is o

Siri’s AI overhaul may have grabbed the headlines at WWDC, but some of Apple’s most useful AI features are arriving elsewhere in iOS 27.

The gap nobody's really tracking Your Git history can tell you that a workstation pushed a commit. What it can't tell you is who or whatactually produced the change. Was it you? An AI agent running inside your IDE? A CI job? Some vendor tool you forgot you'd wired in? For a long time that question was academic. It isn't anymore. The more code we write with AI in the loop, the shakier one quiet assumption gets: that there's a human author behind every commit. Audit trails, incident reviews, compliance workflows; they all lean on it. And it's breaking. Matrix Scroll is a small, open attempt to fix that. It attaches a signed provenance envelope to a commit, and anyone can verify it offline. What it actually does An agent-assisted commit can carry a signed JSON envelope that records: the actor (human or agent) the tool that produced the change an optional bounded scope an Ed25519 signature over a canonicalized version of the manifest The signing input is strict and frankly kind of boring — which is the entire point. It has to be reproducible byte-for-byte across implementations, so: the top-level signature block is stripped before signing object keys are sorted recursively compact separators, ASCII escaping, UTF-8 bytes no NaN, no Infinity The device ID comes from the first eight uppercase hex characters of SHA-256(public_key), formatted as MS-XXXX-XXXX. Verifying is the easy part: take the canonical manifest bytes, check them against the embedded public key and signature. No central service in the middle. Try it without installing anything There's a browser verifier that runs entirely client-side. Nothing gets uploaded: (̿▀̿‿ ̿▀̿ ̿) : https://matrixscroll.com/verify/ Give it ten seconds: Hit Load Commit Envelope → Verify Signature . You'll get VALID, plus the device ID, mode, algorithm, and canonical byte count. Now hit Tamper Sample → Verify Signature again. It flips to INVALID and tells you exactly what broke — e.g. "Device ID mismatch: expected MS-4319-20D5, manifes