DC Studios announces that Director John Joseph Galluppi will helm a new Hellblazer film. This marks the first major creative appointment for the upcoming supernatural project, shifting focus from past attempts.
Source: Headtopics
The global steelmaker recorded third quarter charges of more than $1.1 billion related to the idling of a steelmaking complex in Ukraine that was hit by Russian strikes.
Source: Recycling Today
AppLovin (NASDAQ:APP – Get Free Report) issued its quarterly earnings data on Wednesday. The company reported $4.30 EPS for the quarter, beating the consensus estimate of $4.03 by $0.27, FiscalAI reports. The business had revenue of $2.18 billion during the quarter, compared to the consensus estimate of $2.07 billion. AppLovin had a net margin of […]
Source: Markets Daily
Inseego (NASDAQ:INSG – Get Free Report) announced its earnings results on Wednesday. The technology company reported ($0.27) EPS for the quarter, topping the consensus estimate of ($0.33) by $0.06, FiscalAI reports. The business had revenue of $33.26 million during the quarter, compared to analyst estimates of $31.21 million. Here are the key takeaways from Inseego’s […]
Source: The Lincolnian Online
Episode 2x06, Can't Stop the Healing: Martin has to debunk miracles as homeopath Savannah returns from being jail because of the whole plague thing. He sees Louisa with Parker. A food critic is visiting The Salty Breeze but George isn't himself. Preview for Episode 2x07 is behind the cut: "Love is a Battlefield": Port Wenn has a Revolutionary War reenactment because of course they do. Some of the citizens start displaying tremors and odder-than-usual behavior, so Martin investigates. And Martin is affected by Louisa's new relationship with Parker. I think this show has been steadily improving and I'm a sucker for sad people romance.
Source: Livejournal: Discover Global Communities Of Blogge
Both deputies responded to a report from a victim who stated she had been shot and needed medical attention.
Source: Ntd
Backlash to Operation Metro Surge shaped Minnesota's top races and echoed in Democratic wins across the country on Election Day.
Source: Fox 9 News
According to the American Psychiatric Association, about 1 in 20 Americans are affected by Seasonal Affective Disorder (SAD)—a type of depression linked to a recurring seasonal pattern. In Alaska, that rate is even higher, with roughly 1 in every 10 residents impacted.
Source: Webcenter Fairbanks
It Ends With Booze. Blake Lively launched her non-alcoholic drink line, Betty Buzz, in 2021, and then expanded into alcoholic drinks with Betty Booze in 2023. But TMZ reports that she is seemingly shutting down both brands. TMZ noted that the Betty Booze and Betty Buzz's trademarks are still active, but multiple others related trademark applications to the company have been abandoned. That includes applications for beer-based products, non-alcoholic carbonated beverages and other products. They also note that Betty Buzz hasn't posted on Instagram since April 5, and Betty Booze hasn't posted since September 28. Whole Foods carried the non-alcoholic line, while Total Wine carried the alcoholic line. Source
Source: Livejournal: Discover Global Communities Of Blogge
A closer look at how these manufacturers are advancing thermoforming through automation, material efficiency and application-focused production.
Source: Ein Presswire
As Democrats prepare to take control of both chambers of Congress, experts caution that turning campaign promises into tangible improvements for voters may prove challenging.
Source: Wxii
Several starters were back in action for the Indiana Pacers on Wednesday night. Ivica Zubac returned to the starting lineup after missing the last five games wi
Source: Headtopics
LOST CREEK, W. Va. - Across nine 2026 regular-season sets, the South Harrison Hawks girls volleyball team mostly dominated their head-to-head matchups against the Petersburg Vikings.
Source: Wv News
Two-time All-MIAA guard Jack Bachelor is back with the Ichabods because he's loyal to both the places and people who have made him.
Source: Https://www.wibw.com
Bitcoin price analysis: Sellers failed below $76,000. Now buyers face $81,460 Bitcoin CME futures swept to $75,795 and rebounded on unusually heavy trading. The four-hour recovery is encouraging, but holding $79,040 and clearing $81,460 would tell us much more about whether buyers have genuinely regained control. Bitcoin CME futures have changed character in the space of a few sessions. After aggressive selling and repeated attempts to extend the decline, the market briefly broke below $76,000, recovered, and then rallied to $81,460 on November 4. It is a promising short-term reversal attempt, but the difference between a rebound and a lasting change of trend still matters. The starting point is straightforward: sellers achieved a fresh low but could not sustain the breakdown. Buyers then produced a much more decisive upward move. That echoes the lesson in Greg Michalowski's earlier Bitcoin analysis of sellers failing to extend a support break. His article examines different Bitcoin price references, so the levels in this CME futures analysis should be evaluated on their own terms. How Bitcoin went from heavy selling to a possible reversal October 29: Sellers set the tone. CME Bitcoin futures traded as high as $82,270 before reversing. One four-hour candle fell from an $81,525 open to a $79,045 close. Another ended at $77,025 on volume of 4,141 contracts. There was no need to guess who had the advantage then: the price was moving substantially lower. October 30: Trading activity remained intense, but downside progress became less convincing. A four-hour candle traded between $76,550 and $78,200 on 4,576 contracts, yet opened at $77,425 and closed at $77,420, a net loss of just $5. That does not prove buyers were accumulating Bitcoin. It does show a great deal of trading without a meaningful change between the open and close. November 3: The break below the lows failed. A later four-hour candle fell to $75,795, undercutting earlier lows around $76,260-$76,450, then closed back at $77,905. This could have been a liquidity sweep, when a price move through a well-watched low triggers orders but fails to establish a new lower trading range. We can see the failed break in the price data; we cannot independently verify the stop orders or intentions behind it. November 4: Buyers finally generated meaningful upward displacement. A four-hour candle opened at $78,535, climbed to $81,460 and closed at $80,655, with volume of 5,468 contracts. That was about 5.1 times the preceding 20-bar average of 1,066 contracts. The comparison is striking, although the average includes thinly traded periods around the weekend, so the multiple should not be treated as a calibrated signal. By the last completed four-hour bar in the supplied November 5 snapshot, futures had closed at $80,925. The following bar was still developing. These are dated CME continuous-futures observations, not a live Bitcoin spot quote. The lesson in volume: Effort is useful, but the result matters more Volume measures how much trading took place. It does not, by itself, tell us which side is winning. Compare the two unusually active four-hour candles. On October 30, 4,576 contracts accompanied a candle body of minus $5. On November 4, 5,468 contracts accompanied a candle body of plus $2,120. There was plenty of activity on both occasions, but only the second produced a decisive net change in price. The first pattern is consistent with buyers meeting aggressive selling or with two-sided trading. The second is stronger evidence that buyers were able to push prices higher. Neither observation proves institutional accumulation, trapped short positions or a particular order-flow mechanism. Four-hour open, high, low, close and volume data simply cannot settle those questions. This is why the sequence matters more than an isolated green candle. Heavy selling met limited progress near the lows, a lower break failed, and a subsequent high-volume rally reclaimed resistance. For the November 5 snapshot, that supports a short-term directional assessment of +6 on a -10 to +10 scale. It is a conditional assessment, not a 60% probability, validated trading signal or promise of further upside. $79,040 is the pivot. $81,460 is the next hurdle. The $79,040 region repeatedly interrupted recovery attempts before November 4. The bullish move carried futures back above it. That makes the former resistance area particularly informative on any pullback. A breakout is not the same as a successful retest. Price can rise through a ceiling and immediately fall back below it. A more convincing transition occurs when buyers defend the old ceiling as a new floor, or when price builds acceptance above it without needing a deep pullback. The first nearer-term potential support zone is $80,160-$80,465, drawn from the post-breakout price action. Below it, $79,040 is the more important structural pivot. Overhead, $81,460 is the breakout-session high, followed by the earlier $82,270 high. Reaching a level is not the same as holding beyond it. Why the macro backdrop can help, or interrupt, this Bitcoin recovery The chart gives us a price-based argument, not a complete explanation of why Bitcoin moved. Other markets were also repricing political and economic risks following the US midterm elections. In commodities, oil prices dropped sharply as markets priced hopes of an end to the Iran war. That is a market expectation, not evidence that a diplomatic deal has been completed. Gold's rise alongside falling oil and a weaker dollar offers another useful piece of context: lower energy prices can reduce inflation pressure, but the interest-rate response and the durability of diplomacy remain uncertain. Inside crypto, the November 3 Ethereum outlook emphasized the difficult pre-election and geopolitical backdrop. A November 4 analysis of Zcash's rebound explored the possible influence of falling Treasury yields and a weaker dollar, while also flagging coin-specific factors. Together, these reports suggest that crypto's recovery deserves to be tested against the dollar, yields and broader risk appetite rather than attributed only to technical levels. Yet this is not a uniformly bullish backdrop for all risk assets. The market's positive response to the prospect of slower hyperscaler data-center spending shows that investors can reward some large technology companies while turning away from parts of the AI infrastructure trade. Bitcoin may benefit from improved financial conditions without every equity segment rising alongside it. Three Bitcoin scenarios worth watching 1. Bullish continuation: Buyers clear $81,460. A completed four-hour close above $81,460, followed by sustained trading above it, would make $82,270 the next upside reference. Acceptance above $82,270 would offer stronger evidence of structural recovery. A brief intrabar spike through either level, followed by rejection, would be less persuasive. 2. Constructive consolidation: Buyers defend the breakout. A pullback toward $80,160-$80,465, or even toward $79,040, would not automatically destroy the bullish case. What matters is whether selling loses momentum and buyers can establish support. A successful defense would improve the quality of the recovery, even without an immediate new high. A pullback is a possibility, not a forecast that orders placed there will fill. 3. Failed breakout: Price loses $79,040. A decisive four-hour move back below that pivot, especially if a recovery attempt fails and downside volume expands, would undermine the short-term bullish assessment. The next historical reference is around $77,417, followed by the $75,795 reversal low. A sustained move beneath the latter would seriously challenge the idea that the November 3 break was a durable bottom. When might waiting be more useful than trading? If Bitcoin remains between nearby support and $81,460 without decisive follow-through, the reward for guessing the next move may be limited. More informative observations would include a completed breakout close, a well-defended pullback, or a clear failed reclaim of $79,040. Those are conditions to evaluate, not mandatory entries or a complete trade plan. The takeaway: A failed breakdown is an opening, not a guarantee The Bitcoin futures recovery deserves attention because sellers pushed below prior lows but could not maintain control, and buyers answered with a strong, high-volume move. The important next question is whether Bitcoin can hold the ground it recovered and move above $81,460-$82,270. If it cannot, the rally may remain only a temporary repair within a fragile broader market. The wider-market lesson also matters. In September 1 trading in Japan, 168 of the Nikkei's 225 stocks were rising at one point even as the index slipped and the broader Topix gained. Influential AI names were weighing on the Nikkei despite wider participation elsewhere. A headline index can conceal mixed underlying behavior, just as a Bitcoin rebound can conceal unresolved resistance. Watch what actually advances, what fails to follow through, and whether the market confirms its first move. The next chapter for Bitcoin is not settled by one bullish candle. The response to $79,040 on weakness and $81,460 on strength is likely to be more informative than a prediction made in advance. Bitcoin's $82,000 price magnet: Why reaching the level is only half the story Bitcoin's strong rebound from around $76,000 has brought an interesting question into focus: Could $82,000 become the next destination for buyers? In my view, the $82,000 area is becoming increasingly interesting, not because Bitcoin is guaranteed to reach it, but because several historical price references cluster around $81,900-$82,000. In trading, we sometimes call this a price magnet. Think of it as an area where buyers and sellers may have unfinished business. Previous trading activity can make certain prices particularly interesting, attracting traders looking to enter, exit or take profits. But here's an important lesson: A price magnet is not necessarily a level that price will break through. What should Bitcoin traders watch? The November 4 BTCUSD chart analysis identified several useful areas to monitor: First, $80,750-$80,900. If Bitcoin can climb through this area and remain above it, that would strengthen the case for another advance. A brief jump above resistance is less convincing than price staying there. Next, $81,850-$81,975. This is the main area of interest around the $82,000 price magnet. Traders who bought earlier may consider taking profits, while other participants may be preparing for a breakout. Above $82,050-$82,150. If Bitcoin moves beyond $82,000 and continues trading above this area, it would provide more convincing evidence that buyers are maintaining control. And what if Bitcoin starts falling instead? Holding $80,000-$80,150 would help preserve the recovery. A sustained decline below $79,900 would make me more cautious, with $79,122 becoming another downside reference. These levels come from a November 4 BTCUSD chart snapshot. They are not live prices or interchangeable with CME Bitcoin futures execution levels. Three things can happen when Bitcoin approaches $82,000 1. Bitcoin breaks through. Buyers push above $82,000, and price continues trading there. Traders call this price acceptance. It suggests the market is becoming comfortable at higher prices. 2. Bitcoin breaks through, then reverses. Price briefly rises above $82,000 but falls back below it. This may happen as breakout orders are triggered and sellers respond. It can look bullish initially but turn into a failed breakout. 3. Bitcoin never quite reaches $82,000. Some traders may take profits before the expected resistance, causing the rally to lose momentum. A price magnet can attract interest without ever being touched. Why patience may be the better opportunity I am less interested in chasing Bitcoin after its sharp recovery and more interested in watching how the market behaves as it approaches $82,000. For traders already holding a profitable position, approaching an important resistance area may be an opportunity to consider partial profits or reduce risk. For those looking for a new entry, waiting for a confirmed breakout or a successful pullback may provide more useful information than buying simply because the price is rising. The broader lesson applies far beyond Bitcoin: Predicting where price might go is only one part of trading. Understanding how price behaves when it gets there can be even more valuable. The destination is interesting. The reaction at the destination may be even more important. Educational only. Trade at your own risk. This article was written by Itai Levitan at investinglive.com.
Source: Forexlive
Las Vegas City Council voted to let Arts District venue Cork and Thorn stay open under strict conditions, but the owner faces eviction.
Source: Https://www.fox5vegas.com
Oklahoma City Thunder guard Shai Gilgeous-Alexander scored 42 points and dished out eight assists in a 124-116 win over the Cleveland Cavaliers on Wednesday night.
Source: Yardbarker
Haddonfield defeats Lindenwold - Boys soccer recap
Source: Nj
Miner in demand from ‘long’ investors, according to UBS
Source: Mining Journal
Geometric Deep Learning: Mapping non-Euclidean manifolds, graphs, and 3D molecular structures beyond flat grid representations. Why standard Euclidean neural networks fail in non-flat domains—and how graph manifolds are revolutionizing molecular biology, 3D computer vision, and network science. For over a decade, deep learning achieved phenomenal success by operating on flat, regular grid structures. Convolutional Neural Networks (CNNs) conquered two-dimensional pixel arrays in images, while Transformers excelled at one-dimensional sequential token streams in text. However, the real world rarely fits into neat 1D lines or 2D grids. Molecules, social networks, protein folding structures, 3D point clouds in autonomous driving, and financial transaction graphs are fundamentally non-Euclidean. Forcing these complex, irregular structures into flat matrices discards critical spatial, topological, and geometric invariants. To overcome these structural limitations, researchers and domain specialists are embracing Geometric Deep Learning (GDL). What is Geometric Deep Learning? Geometric Deep Learning is a unified mathematical framework that generalizes deep learning architectures to non-Euclidean domains—such as graphs, manifolds, and 3D geometric meshes. Rather than assuming fixed grid coordinates, GDL builds neural operations around underlying mathematical symmetries, permutation invariances, and geometric transformations. At its core, GDL bridges two major structural representations: Graph Neural Networks (GNNs): Designed for irregular, relational data structures consisting of arbitrary nodes and interconnecting edges. Manifold & Equivariant Networks: Designed for continuous curved surfaces, 3D spatial points, and physical symmetry groups where transformations (like rotation or translation) must preserve spatial meaning. Core Operational Mechanics: Message Passing and Equivariance Two fundamental concepts enable Geometric Deep Learning to process complex geometric domains: Neural Message Passing In Graph Neural Networks, feature aggregation happens through a localized "message passing" paradigm. Each node collects feature vectors from its immediate neighbors, aggregates them using permutation-invariant functions, and updates its internal state. This allows the network to learn topological patterns regardless of graph size or node ordering. Geometric Equivariance and Invariance In physical domains (like molecular dynamics or 3D computer vision), rotating a 3D protein structure shouldn't change its predicted binding energy (invariance), but rotating a 3D point cloud should predictably rotate its predicted bounding box (equivariance). GDL bakes these physical symmetries directly into the model's weight matrices, preventing the network from having to relearn physics from scratch. Why Non-Euclidean Architectures Matter Generalizing deep learning to irregular geometries unlocks major technological capabilities: High Sample Efficiency in Physical Sciences By enforcing physical symmetries and geometric laws directly inside the architecture, GDL models require exponentially fewer training samples to predict molecular properties, protein folds, or material behaviors accurately. Relational Intelligence over Heterogeneous Networks Traditional neural nets treat data points independently. GDL models natively capture complex, multi-relational dependencies across massive connected networks, such as financial transaction graphs or multi-protein interaction networks. Native Processing of 3D Spatial Sensor Data Autonomous vehicles and spatial computing hardware capture the world as irregular 3D point clouds (via LiDAR and depth sensors). GDL operates natively on point cloud manifolds without losing precision through spatial voxelization. High-Impact Real-World Applications Generative Drug Discovery and Materials Science: Designing novel therapeutic molecules and crystalline energy materials by predicting 3D atomic bond interactions natively. 3D Vision and Autonomous Driving: Processing real-time LiDAR point clouds for high-precision spatial tracking and obstacle detection in unstructured environments. Financial Anti-Money Laundering (AML): Analyzing enterprise transaction graphs in real time to spot complex, multi-account money laundering cycles. The Next Era of Spatial and Physical Intelligence As artificial intelligence expands deeper into the physical sciences, spatial computing, and complex network systems, flat grid representations are no longer sufficient. Geometric Deep Learning provides the mathematical foundation needed to extend deep learning across the true, non-Euclidean geometry of the physical world.
Source: Dev Community
Commuters wait to board a train at Central railway station in Sydney, Australia, on Jan.15, 2025. (Lisa Maree Williams/Getty Images)
Source: The Epoch Times
Former deputy was convicted in July in separate case for tipping off drug dealers to federal probe
Source: Albuquerque Journal
NNN REIT (NYSE:NNN – Get Free Report) released its quarterly earnings data on Wednesday. The real estate investment trust reported $0.57 earnings per share (EPS) for the quarter, beating analysts’ consensus estimates of $0.52 by $0.05, FiscalAI reports. NNN REIT had a return on equity of 8.70% and a net margin of 40.35%. The business […]
Source: Ticker Report
Recently, a fan's post about why she was unstanning her bias, SS501's Heo Youngsaeng, went viral, triggering a huge backlash from netizens.
Source: Koreaboo
Danang Plus is a small independent guide to Da Nang, Vietnam. It has two editions: Vietnamese at danang.plus and English at en.danang.plus. The English edition is a static site built with Astro. This post collects a few practical notes from keeping two language editions of the same guide in shape, in case they help someone planning a similar project. Why two subdomains instead of /en/ folders Both options work. Subdomains made sense here because the two editions are written separately rather than machine-translated, they have different editorial priorities, and they could be deployed independently. The trade-off is that each subdomain is treated as its own property in search tools, so you verify and monitor both, and you submit two sitemaps. Write for the reader, not as a translation A visitor reading in English usually needs different information from a local reader. An English page about renting in Da Nang has to explain deposits, contracts and temporary residence registration, things a Vietnamese reader already knows. A Vietnamese page on the same topic can skip that and go straight to neighbourhoods and prices. Treating each edition as its own content plan, rather than a one-to-one copy, produced more useful pages on both sides. Keep facts that change in one place Prices, opening hours, ticket costs and transport fares go out of date quickly. Pages such as the cost of living in Da Nang list where each figure came from and when it was checked. Keeping a "last checked" date next to volatile facts makes it obvious which pages need review first, and it tells readers how much to trust a number. Sitemaps per content type The English sitemap index splits into posts, pages, categories and authors. That makes it easy to see in a search console which type of page is or is not getting indexed, and to re-submit only the part that changed. Small things that pay off Use one canonical URL per page and stick to a trailing-slash convention. Link related guides to each other in the text, for example from a beach guide to where to stay in Da Nang, so readers and crawlers can move through the site. Keep images compressed and sized for the layout; a static build makes this easy to automate. Put the date a guide was last updated near the top. Wrapping up None of this is complicated, but writing each language edition for its own readers, and making the date on every changing fact visible, did more for the guide than any single technical trick. If you are curious how it turned out, the English edition is at en.danang.plus.
Source: Dev Community
using registers an object with a disposal method in a scope and calls cleanup when that scope ends. When several resources were registered, they were disposed in reverse acquisition order. An early return followed the same rule. Nine conditions in Node.js covered disposal order, retained exceptions, and asynchronous waiting. Before using it, check which object supplies disposal and which scope exit triggers cleanup. Disposal finishes before the caller receives a return value This code ran as an .mjs file in Node.js 24.15.0, without a transpiler or polyfill: The result contained value: "done", with events open:A → open:B → body → close:B → close:A. Both disposals had finished when the caller received the return value. open only returns a fixture object. The log records calls to Symbol.dispose; it is not a measurement of real files or connections. The ECMAScript declaration and scope specification also describes disposing resources registered through using when scope processing finishes. The comparison included failures as well as normal completion: Condition Observed order or result Acquire A and B, finish normally Dispose B, then A Acquire A and return Dispose A before returning to the caller Throw in the body Dispose A before entering catch Acquire A, then throw while acquiring B Dispose registered A; no disposal call for B Body and both A/B disposals throw Attempt both disposals and retain errors through SuppressedError Register A and B with await using Finish B's asynchronous disposal before disposing A Pass an asyncDispose-only value to using TypeError; asynchronous disposal is not called Pass null and undefined to using Execute the body without disposal methods Throw while using a temporary directory Directory no longer exists after entering catch When acquiring B throws, using does not dispose B afterward. If the acquisition function allocated something before failing, that function needs to clean up its partial acquisition. Where does the body error go if cleanup fails too? A deliberately failing disposer tested this case. The body threw body failed, B threw close failed:B, and A threw close failed:A. The caught exception's error.message was close failed:A. The body failure was still retained: Reference Recorded message error.error.message close failed:A error.suppressed.error.message close failed:B error.suppressed.suppressed.message body failed Both the outer and inner wrappers were SuppressedError objects. Cleanup continued after a disposal failed, with earlier errors retained in a related chain. The explicit resource management design material explains this relationship. Logging only the outer message can miss the body failure. The experiment followed both error and suppressed and checked each message. Use await using for asynchronous disposal The asynchronous fixture recorded the start and end of Symbol.asyncDispose separately. Registering await using a = ... and await using b = ... produced body → close start:B → close end:B → close start:A → close end:A → after work. The wait in this fixture was Promise.resolve(). It did not measure network or disk I/O duration. It did verify that A's disposal started after B's completed, and that the caller continued after both finished. Passing an object with only Symbol.asyncDispose to using without await produced a TypeError. An asynchronously disposable value cannot simply be placed in an ordinary using declaration. Check a real Node.js temporary directory The next example uses an actual temporary directory rather than only a logging object: import { existsSync, mkdtempDisposableSync, writeFileSync } from 'node:fs'; import { tmpdir } from 'node:os'; import { join } from 'node:path'; let path; catch (error) { } The output was inside: true, caught: work failed, and after: false. The directory, including sample.txt, was removed when the exception left the scope. Node.js 24.15.0's mkdtempDisposableSync returns an object with a path and a disposal method for removal. This API was added in 24.4.0. An arbitrary path string does not acquire that deletion behavior merely by being used in a using declaration. The tests ran on September 11, 2026, with Node.js 24.15.0 on macOS arm64. The nine-condition fixture and results are available. Other Node versions, browsers, transformed code, and forced process termination were not tested. These results cover disposal when JavaScript can execute the scope-exit steps.
Source: Dev Community
This review contains spoilers. Manual Cinema’s “Frankenstein” began with a literal bang: lightning crashing across the screen, accompanied by sounds of thunder and flickering lights. Directed by Drew Dir, the production ran at Chicago’s Studebaker Theater Oct. 28 through Nov. 1, retelling the Mary Shelley classic through shadow puppetry, live music and other creative visual... The post Curtain Call: Manual Cinema’s ‘Frankenstein’ revives old story with shadow puppets, live music appeared first on The Daily Northwestern.
Source: The Daily Northwestern
Brooks Nader is hopping into bunny territory ... and it looks like she's the newest addition to Hollywood's upcoming Playboy movie, "Playmates!" Photos of the "Baywatch" babe rocking the iconic Playboy Bunny look while filming on set at a Beverly…
Source: Tmz
Growing competition for semiconductor talent is fueling a battle over trade secrets as Chinese chipmakers gain ground.
Source: Cnbc
On 25 October 2026, a developer reported that updating Postman to version 12.33.4 silently installed Postman plugins into Claude Code and Codex, and registered its startup hook as trusted. It matters because a coding agent runs with your credentials and codebase, so anything that changes what it loads changes what it can do for you. Key takeaways Postman 12.33.4 allegedly installed its own plugins into Claude Code and Codex during an ordinary app update, without an explicit opt-in step. The report says a Postman SessionStart hook was added to the agent's trusted list, so it runs before every session and steers API work towards Postman's skill. A two-minute audit of your settings files and plugin directories shows whether your machine is affected. Removal is straightforward: delete the plugin entry, the marketplace entry and the trusted hook record, then restart the agent. No evidence in the original issue shows API data being shared — that stronger claim circulates on social media but is not documented in the report itself. What was actually alleged? The allegation comes from a GitHub issue filed on 25 October 2026 by a developer using the handle indicava, titled with a severity marker and still open at the time of writing, with 453 reactions and 51 comments. A Hacker News thread on the same report reached 1,391 points and 461 comments, which is why the story travelled far beyond the people who use Postman daily. According to the issue, the installer's own log records the event with the source field set to "auto-existing-upgrade", followed by claude-code:installed and codex:installed. In plain terms, the upgrade path itself triggered the plugin installs, not a separate action by the user. The reporter then found three concrete changes on their machine: Their ~/.codex/config.toml had gained a Postman plugin marketplace entry. That marketplace had an enabled postman@postman plugin. The hooks state had gained a trusted_hash entry for Postman's hook, which is the record that lets a hook run without asking again. The hook in question is a SessionStart hook, and the report says it instructs the agent to load the postman:api-engineer skill before any non-trivial API engineering task. The issue also points to an approvePostmanHook step in the installer source, which appears to approve Postman's own hook programmatically rather than waiting for the user to approve it. None of this has been independently reproduced in public at the time of writing, and the report describes one developer's machine. Treat it as a credible, specific allegation with receipts in the form of log lines and config diffs — not as a confirmed finding about every Postman installation. Why does a quiet plugin install matter so much? A plugin for an ordinary editor is a convenience. A plugin for a coding agent is a change to the agent's behaviour, and that distinction is the whole story. First, a SessionStart hook runs before you type anything. By the time you see your first prompt, the agent has already loaded instructions you never reviewed. Second, a skill that loads "before any non-trivial API engineering task" does not just add a feature; it steers every API decision the agent makes towards one vendor's tooling. Third, the trusted record exists precisely to stop silent changes: the confirmation dialog is the one moment a user gets to say no, and writing your own entry into the trusted list removes that moment entirely. As one commenter put it, the precedent is the problem: if a mainstream desktop app can change how your coding agent behaves during a routine update, then the trust boundary of the agent is whatever any installed app decides it is. That is a much bigger claim than "Postman added a plugin", and it is worth taking seriously even if you never use Postman. How do I check my own machine? You do not need to install anything to audit this. The check is read-only: it looks, it does not change. Run it in your terminal: # 1. Claude Code: settings files and plugin directories ls ~/.claude/plugins/ 2>/dev/null grep -ri "postman" ~/.claude/settings.json ~/.claude/settings.local.json 2>/dev/null # 2. Codex: the config file named in the report grep -n -i "postman" ~/.codex/config.toml 2>/dev/null # 3. Sweep both tool directories for any Postman mention at all grep -rn -i "postman" ~/.claude/ ~/.codex/ \ --include="*.json" --include="*.toml" 2>/dev/null | head -40 What the output means: no matches means no trace of Postman in either tool's configuration, and you can stop here. Matches in config.toml that mention a marketplace, a postman@postman plugin or a trusted_hash entry match the pattern described in the report. Matches elsewhere, such as an MCP server entry you added yourself, are a different thing — a plugin you installed deliberately is not the problem this article is about. One honest limitation: this sweep covers the two tools named in the report and their usual config locations. If you run other agents, their plugin directories deserve the same five minutes, because the same technique would work there too. What do I do if I find it? Back up first, then remove the three pieces. The backup costs nothing and makes every later step reversible: cp ~/.codex/config.toml ~/.codex/config.toml.bak cp ~/.claude/settings.json ~/.claude/settings.json.bak 2>/dev/null Open ~/.codex/config.toml and delete the Postman marketplace entry, the enabled postman@postman plugin line, and the trusted_hash entry under the hooks state. In Claude Code, list installed plugins and uninstall the Postman one, then check ~/.claude/settings.json for any Postman hook entry and remove it. Restart both agents completely, so no already-loaded session keeps the old configuration in memory. Re-run the audit commands above and confirm the output is empty. Do not stop at step 4. Open Postman itself and check whether its settings offer an option for the CLI or agent plugins, and switch it off if it does — otherwise the next update may simply reinstall what you just removed. If no such option exists, that absence is itself useful information to include in any bug report you file. What has Postman said, and what is still unproven? A reply in the GitHub issue, from an account that states no affiliation with Postman, says the CLI and agent plugins are bundled because users struggled to install them separately, that the bundling is noted on the download page and in the update flow, and that the plugins can be uninstalled from inside the app. The same reply says the team is going to make the opt-in clearer, especially for existing users, and is reviewing the hook approval and session-start behaviour. That is a meaningful concession on the opt-in question, but note what it is not: it is not a verified statement from Postman, and it does not address whether writing a trusted hook record was intended. Equally important is what nobody has shown. The original issue contains no evidence that API keys, request bodies or production tokens were sent anywhere. The stronger version of the story — that Postman "shares all API data" — appears in social media threads and video explainers, but the details behind it are not spelled out in the issue itself. Repeating the stronger claim as fact would be exactly the kind of sloppy reporting this site tries to avoid. Similarly unproven: how many users were affected, whether the behaviour was a deliberate growth tactic or a packaging bug, and whether it affects every platform or only the reported setup. How do I stop this happening again? Audit after every update. The two-minute sweep above is worth running after any update to an app that integrates with your coding tools, not just Postman. Put agent settings under version control. A git repo for ~/.claude and ~/.codex config turns a silent change into a visible diff. You will see the next surprise the day it happens. Treat hook approvals as security decisions. Never approve a hook you did not ask for, and periodically review the trusted list the same way you review SSH keys. Keep agent API keys scoped. A key that can only read a staging environment limits the damage if an agent ever does something you did not authorise. Prefer explicit installs. If a tool's plugin is worth having, installing it yourself takes thirty seconds and leaves no doubt about consent. The deeper lesson is not really about Postman. Coding agents collapse the distance between "software installed on your machine" and "software acting on your behalf". Until agent tools treat plugin installs and hook approvals with the same seriousness as browser extensions or OS permissions, the audit above is part of basic hygiene — the equivalent of checking what is running at startup. Run it now, while the story is fresh, and keep it in your routine after that. Originally published on LogicStack. Run the code examples right in your browser with our free interactive Python and SQL playgrounds — no setup, no sign-up.
Source: Dev Community
Anthony Federline appeared in court Wednesday on charges alleging he sexually assaulted his 12-year-old stepdaughter.
Source: Necn
Officials will conduct a full recount after approximately 100 printed paper ballots were fed into a voting tabulator despite a requirement that they be counted…
Source: Necn
Boston Public Schools faces scrutiny as students and parents push back on a plan to close Charlestown High School. Meetings are set Thursday night.
Source: Necn
DC Studios just made a major move for its next supernatural project, and it’s not the sequel fans have been waiting for. A new John Constantine movie has hired its first key creative, signaling a fresh direction for the iconic Hellblazer. Evil Dead Wrath writer to pen new John Constantine reboot DC Studios has tapped […] The post New John Constantine Movie Hires Its First Name & It’s Not Keanu Reeves appeared first on ComingSoon.net - Movie Trailers, TV & Streaming News, and More.
Source: Coming Soon
OpenAI has rolled out ChatGPT Images 2.6 across consumer tiers, Codex, and its developer platform, introducing native animated GIF creation. Read more...
Source: Neowin Net
EL PASO COUNTY, Colo. (KRDO) -- Ballot Issue 7A, the measure about the Front Range Passenger Rail District's (FRPRD) project, "CoCo" passed statewide on Election Night. The CoCo team says nine of the 12 counties voted in favor of funding the project by establishing a sales use tax that will cost consumers one-third of a The post What’s next for CoCo after Ballot Issue 7A passes appeared first on KRDO.
Source: Krdo
I work with AI agents every day, and the hardest problem I've run into isn't model quality. It's credentials. The agent needs to log in to something to finish the job, and the default answer is pasting a password into chat: the secret ends up in a transcript and a model's context, unscoped and unrevocable in practice. This post is about the design principles I've landed on for solving that, and how I'm applying them in AgentKey, the agent credential vault I'm building. I want feedback on the principles more than the product. Principle 1: structural separation, not behavioral trust The model should never need to hold the plaintext credential. Not "we trust it not to look" and not "we redact afterwards." The secret and the model live on different sides of a boundary, and only a narrow execution layer may cross it, after explicit owner approval. If your design relies on the model choosing not to misuse a secret it can see, you don't have a security boundary; you have a policy hope. Principle 2: the human's job is one decision Every extra step between "agent asks" and "human approves" pushes people back to pasting secrets into chat. The approval has to be one tap, in the place the human already is. In AgentKey that's a request link: the agent sends a link, the owner opens it, approves on the phone, done. Approving once grants ongoing use of that login until revoked. The link can lapse; the approval shouldn't silently expire on a timer the owner never set. Principle 3: scope per agent, per credential, with receipts A grant is never "the agent may use my accounts." It's per login, or explicitly all-access, per agent. Revocation is a single switch, and revoked access goes to an archive the owner can restore. Requests and approvals are recorded in an activity view. If the owner can't answer "who can use this password right now?" in one glance, the model is wrong. Principle 4: agents are contacts, not infrastructure If connecting an agent feels like provisioning, the workaround wins. In AgentKey, agents are contacts you add like people. Agents that live inside other apps connect through standard OAuth consent as custom connectors; Claude and ChatGPT both support that with a name and a URL. The connector's vocabulary is request links and statuses. It never carries secrets. Principle 5: execution is a separate, narrow layer Approval says "may you." Performing the login is a different job. The principle that follows is separation: one dedicated executor - a sandboxed browser session - is the only component meant to touch the plaintext, while the agent works against a scrubbed view of the page, labels rather than values. An AI in that loop reads the login page instead of hand-maintained per-site adapters, because a credential vault that only works on a compatibility list isn't a vault, it's a plugin collection. Worth comparing notes If you're working on agent identity or permissions, I'd genuinely like to compare notes - for example on where task scoping should live (in the grant or in the executor), how agents should prove which task a credential use belongs to, and what audit format actually works for a human who checks receipts twice a year. Site: https://agentkey.waiwiser.com
Source: Dev Community
Different iterations of the Marvel universe collide in Avengers Doomsday, just as many different flavors of tie-in snacks will arrive (or already have) at fast-food joints and chain restaurants. Many heroes will not survive the coming conflagration; can your digestive system survive the many forms of Doctor Doom-themed food, especially when it’s loaded with green […]
Source: Comic Book
This is a submission for the Presage SmartSpectra SDK Challenge What I Built Hyperion Core: Building a Transparent Seismic Research Dashboard Earthquake information is important—and the way we present it matters. With Hyperion Core, I’m exploring how a software prototype can bring together public earthquake catalog data, interactive visualizations, and small training demos while being clear about what is real and what is simulated. Hyperion Core is a Japan-focused, bilingual research and training dashboard. It fetches recent earthquake events from the USGS past-hour catalog, filters them to a broad Japan-region area, and displays details such as magnitude, depth, location, and event time. The dashboard also includes a Japan-centered globe, a prefecture directory, preparedness resources, an offline-capable app shell, and local peer-to-peer messaging demos. The technology behind it The main dashboard is built with Next.js, React, and TypeScript. JavaScript, HTML, and CSS support browser interactions and interface styling. The repository also contains bounded examples and starter modules in Rust, Zig, C++20, Julia, Python, Haskell, Go, Java, R, PHP, and Swift. These explore ideas such as frame parsing, signal processing, scientific modeling, and inference adapters. They are examples and scaffolds—not a collection of production-ready services. What models does it use? There are two distinct model-related features: Gemini Live API: An optional, user-started English/Japanese voice conversation feature. The configured default model is gemini-3.8-live. It supports conversational interaction; it is not an earthquake prediction model or an official alert channel. A logistic regression demo: A browser-based training lab generates a synthetic dataset and trains a small classifier. It demonstrates a machine-learning workflow, but the data is artificial and the results say nothing about real earthquake risk. Hyperion Core does not include a validated or pretrained seismic prediction model. Its visualized signals and hardware telemetry are illustrative, not measurements. Why be explicit about the limits? Earthquake and emergency information can influence important decisions. This project is not connected to field sensors, Japan’s official Earthquake Early Warning system, emergency dispatch, sirens, or infrastructure controls. Its visuals are demonstrations, and the USGS catalog lists detected events—it is not a real-time warning feed. Please use JMA and local authorities for official information and emergency guidance. Hyperion Core is a research prototype for exploring software and communicating data boundaries, not a tool for operational or emergency decisions. Live demo: https://hyperion-core-secure-edge.aniruddhadhibar5.workers.dev webdev #nextjs #typescript #react #machinelearning #earthquake I based the language and model details on the repository: Gemini Live is configurable, the logistic regression lab uses generated data, and no validated seismic predictor is included. Code <!-- Share a link to your public code repository. Make sure it includes setup instructions. --> https://github.com/aniruddhadhibar5-bot/ApexTectonic How It Responds to Human Signals Prize Categories Consent and Data Handling Agent Session
Source: Dev Community
JAKARTA, Nov 5 (Reuters) - Australia will join Indonesia's waterbombing operations to help put out fires that have been raging for months,...
Source: Usnews