{"list":"front","lang":"en","page":5,"has_more":true,"next":"https://startupxo.com/api/v1/front.json?lang=en&page=6","html":"https://startupxo.com/en/","markdown":"https://startupxo.com/en/index.md","items":[{"id":1225,"kind":"story","title":"User report that the Postman desktop app automatically installed Codex and Claude Code plugins","url":"https://github.com/postmanlabs/postman-app-support/issues/14370","host":"github.com","by":"startupxo","score":0,"comments":0,"time":"2026-11-03T18:25:07.825Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["The issue author reported that after opening the Postman desktop 12.33.4, the Postman log recorded that Codex and Claude Code plugins were installed through an automatic upgrade path.","According to the report, a Postman remote MCP connection and a hook telling Codex to load the Postman skill when a session starts were added to Codex's global settings. The install source code reportedly contains logic to add this hook to the trusted list after installation.","The report states that this is an observation the reporter made on their own Mac and that it was not reproduced in a clean environment. The report does not claim data exfiltration. The original text contains no official response from Postman."],"discussion":"https://startupxo.com/en/1225/","markdown":"https://startupxo.com/en/1225.md","api":"https://startupxo.com/api/v1/item.json?id=1225&lang=en"},{"id":1224,"kind":"story","title":"How to check the format and content of AI-written design records with regomark","url":"https://zenn.dev/babarot/articles/regomark-adr-checker","host":"zenn.dev","by":"startupxo","score":0,"comments":0,"time":"2026-11-03T12:29:03.012Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["An ADR is a document that records design decisions. The author writes that ADRs entrusted to AI gradually developed section structures that drifted, and that decision text came to include code details such as screen menu lists and constants.","The author proposes checking structural rules that a machine can verify, such as section order, Status values, references between ADRs, and whether linked files exist, with a policy file that fails CI on violations. Semantic rules, such as whether a document reads like a specification, are judged by an LLM and left only as warnings on the PR.","regomark is a command-line tool the author built. It analyzes Markdown and checks rules with OPA (a policy evaluation tool). The author states that there is no ADR-specific implementation and that the rules live in the policy file of the repository that uses it.","When the author re-read 7 of their own repositories, they report that everything that differed from the code was detail that did not need to be in the ADR, and that the decisions and guarantees to be kept were not outdated. This is one author's observation."],"discussion":"https://startupxo.com/en/1224/","markdown":"https://startupxo.com/en/1224.md","api":"https://startupxo.com/api/v1/item.json?id=1224&lang=en"},{"id":1223,"kind":"story","title":"A cyberattack changed freezer and refrigeration settings, and European supermarket food was discarded in large amounts","url":"https://www.nikkei.com/article/DGXZQOGN02B1R0S6A101C2000000/","host":"nikkei.com","by":"startupxo","score":0,"comments":0,"time":"2026-11-03T06:33:53.364Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["Belgian supermarket chain Delhaize was hit by a cyberattack between 1 and 2, and the temperatures of refrigeration and freezer units in about 40 stores rose. The source says local media reported that an meat chain in Antwerp, Belgium, was also attacked, and that staff manually operated equipment to lower temperatures.","According to the source, the attack changed equipment settings to defrost mode, which raised temperatures. Attacks using the same method also hit hundreds of large supermarkets in France in mid October, and large amounts of food were thrown away at that time.","The source says similar attacks occurred at several supermarkets in Sweden, about 10 stores in Ireland, and in Estonia. Damage at Delhaize was concentrated in small stores in Belgium and Luxembourg.","The EU commissioner for food safety, Hansen, criticized the attacks as an attempt to deliberately disrupt the food supply system, and Belgian authorities started an investigation. The source does not identify who carried out the attacks."],"discussion":"https://startupxo.com/en/1223/","markdown":"https://startupxo.com/en/1223.md","api":"https://startupxo.com/api/v1/item.json?id=1223&lang=en"},{"id":1222,"kind":"story","title":"Author-measured case where a longer Snowflake clustering key improved Japanese address filter efficiency","url":"https://zenn.dev/snowflakejp/articles/d907c02698d748","host":"zenn.dev","by":"startupxo","score":0,"comments":0,"time":"2026-11-02T18:26:16.606Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["Snowflake stores tables split by sorted clustering column values and reads only the parts that match a condition to speed up searches. The old method used only the first 5 bytes of a string column, but the source says Optima Clustering can use up to 1KB of the whole clustering column combined.","In UTF-8, one Japanese character takes 3 to 4 bytes, so 5 bytes covers only about 1.25 to 1.33 characters. The author explains that addresses starting with '東京都' were barely filtered by ward-level conditions.","The author built CLASSIC and OPTIMA tables from about 19.58 million rows of the MLIT Street-Block Level Location Reference Information CSV. For the LIKE condition '東京都中央区%', the author wrote that CLASSIC read 3 partitions and OPTIMA read 1 partitions.","These figures are direct measurements by the author on one dataset and a few conditions. The author is an individual engineer at Snowflake, and the source gives no evidence that the same results hold for other data and queries."],"discussion":"https://startupxo.com/en/1222/","markdown":"https://startupxo.com/en/1222.md","api":"https://startupxo.com/api/v1/item.json?id=1222&lang=en"},{"id":1221,"kind":"story","title":"Ford Announces 'Privacy Promise' to Explain How It Uses Connected Car Data","url":"https://techcrunch.com/2026/11/02/ford-launches-a-privacy-promise-and-explains-how-and-why-it-uses-a-customers-data/","host":"techcrunch.com","by":"startupxo","score":0,"comments":0,"time":"2026-11-02T12:29:40.268Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["Ford announced its 'Privacy Promise' on November 2. It said it will protect vehicle data and let customers decide how much to share. It also opened a website that explains how data is collected and used and shows how to change settings.","The website will cover the types of data collected, how the data is used, and how to agree to or withdraw from data sharing. Ford's chief privacy officer said the settings feature has existed for years and that this effort aims to explain it more clearly.","Ford said it does not sell vehicle driving behavior data to data brokers and does not use it for targeted advertising. This is the company's claim, and the source did not independently verify it.","According to several surveys cited in the source, only 14% of U.S. consumers said they fully know what data their vehicles collect. Ford's chief privacy officer said Ford's own survey of 1,000 people found that 91% consider data privacy important."],"discussion":"https://startupxo.com/en/1221/","markdown":"https://startupxo.com/en/1221.md","api":"https://startupxo.com/api/v1/item.json?id=1221&lang=en"},{"id":1220,"kind":"story","title":"Apple Creator Studio update adds AI assistants and new effects for video editing and document work","url":"https://images.apple.com/newsroom/2026/10/apple-creator-studio-adds-powerful-new-ai-features-and-even-more-creative-tools/","host":"images.apple.com","by":"startupxo","score":0,"comments":0,"time":"2026-11-02T06:40:54.738Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["Final Cut Pro for Mac includes MotionVFX visual effects, transitions, and templates by default, and Apple said Content Hub stock footage selected by Apple can be found and used directly inside the editing app.","Keynote, Pages, and Numbers get an AI assistant in beta that understands document content. According to the source, Keynote helps with slide design and animation, Pages helps with draft writing and structure, and Numbers helps with chart recommendations and table creation.","Pixelmator Pro's Generate Fill lets you select part of an image, describe in a sentence what you want to replace it with, and fills the area to match the surrounding lighting, perspective, and texture. The claim that it blends naturally is Apple's claim.","Logic Pro for Mac's Stem Splitter was introduced as making it possible to separate, replace, or balance the drum part even without the original multitrack recording. Logic Pro's AI assistant is also available in beta."],"discussion":"https://startupxo.com/en/1220/","markdown":"https://startupxo.com/en/1220.md","api":"https://startupxo.com/api/v1/item.json?id=1220&lang=en"},{"id":1219,"kind":"story","title":"A beginner who used an AI coding agent to join a Japanese web performance contest and raised the score more than 4 times in one day","url":"https://zenn.dev/nicklegr/articles/de0a6fd86b6ce3","host":"zenn.dev","by":"startupxo","score":0,"comments":0,"time":"2026-11-02T00:44:30.380Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["ISUCON is a web service performance contest held in Japan that scores whether an SNS app with photo posting and friend features works accurately and quickly. This article is a record of an author with little infrastructure experience who took part using the AI coding agent Codex.","The author reports that the score rose from 4,645 points on the official load test to a final 19,593 points, about 4.22 times higher. The provisional ranking the author tallied was rank 497.","The largest gain came from adding 4 DB search indexes, which recorded 18,395 points (about 70.5% higher). Signature calculation and DB connection reuse also raised the score, and one change dropped the overall score to 16,623 points, so it was reverted.","A point was raised that INSERTs targeting all users cause long locks in the notification feature. The agent proposed narrowing the lock scope, but the author said that a person approved the change after confirming whether structural changes are allowed under the contest rules."],"discussion":"https://startupxo.com/en/1219/","markdown":"https://startupxo.com/en/1219.md","api":"https://startupxo.com/api/v1/item.json?id=1219&lang=en"},{"id":1218,"kind":"story","title":"Browser data containing login states is traded on black markets, why some people separate banking and work browsers","url":"https://togetter.com/li/2765521","host":"togetter.com","by":"startupxo","score":0,"comments":0,"time":"2026-11-01T18:23:47.127Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["Dark web researcher Taiki Watanabe, whose post was featured on Togetter, claims that the main product on black markets has shifted from lists of IDs and passwords to browser data containing login states and device characteristics.","According to the presenter, this data includes login information such as cookies and device fingerprints like screen size, fonts, and language settings, and buyers can reproduce the same environment with an anti-detect browser.","The original text says personal data is priced at 1 per item, ranging from 300 yen to 1,500 yen, and access data for corporate admin screens sells for tens of thousands to hundreds of thousands of yen, but the source and counting method are not given in the text.","According to the presenter, data can leak even from the latest Chrome through malicious extensions or by interfering with browser processes, and this has not been independently verified."],"discussion":"https://startupxo.com/en/1218/","markdown":"https://startupxo.com/en/1218.md","api":"https://startupxo.com/api/v1/item.json?id=1218&lang=en"},{"id":1217,"kind":"story","title":"How to Give AI Agents a Work Procedure Manual: Design in 8 Ways, Shown Through a Scheduling Case","url":"https://techblog.kazaneya.com/20261030-agent-skills/","host":"techblog.kazaneya.com","by":"startupxo","score":0,"comments":0,"time":"2026-11-01T06:30:26.022Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["Agent Skills are work procedure manuals written so that AI can read and follow them. Using scheduling as an example, the author explains how to write steps that first ask for missing information, then calculate free time, and then get human approval before anything is sent externally.","The author advises leaving processing that has a single correct answer, such as date calculation or conflict checks, to scripts, and having AI handle only judgment and writing. The stated reason is to ensure the same input always produces the same result.","The author's team records the reasons behind its judgment criteria, keeps past failure cases as automated tests, and runs trial operations in which a new AI session with no prior information runs the procedure alone to find gaps. This is the author's team's operating method and has not been verified externally."],"discussion":"https://startupxo.com/en/1217/","markdown":"https://startupxo.com/en/1217.md","api":"https://startupxo.com/api/v1/item.json?id=1217&lang=en"},{"id":1216,"kind":"story","title":"Unfortunate 9 patterns often seen in AI-written code comments","url":"https://qiita.com/jnchito/items/636c519c0b1d347c79b7?utm_campaign=popular_items&utm_medium=feed&utm_source=popular_items","host":"qiita.com","by":"startupxo","score":0,"comments":0,"time":"2026-11-01T00:48:29.059Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["Many comments simply restate what the code already makes clear. The original article recommends removing such comments and writing why something is done only when truly needed.","A comment should sit directly above the code it explains. If a note about a specific step inside a method is placed above the method definition, it is easily mistaken for a description of the whole function.","According to the original, information such as past change history or specific call sites is better placed in commit messages, pull request descriptions, or comments in the calling code than in code comments.","The original author says AI-written comments should not be trusted as-is and should be reviewed by people. They also suggest organizing comment rules into an AI instruction file such as CLAUDE.md."],"discussion":"https://startupxo.com/en/1216/","markdown":"https://startupxo.com/en/1216.md","api":"https://startupxo.com/api/v1/item.json?id=1216&lang=en"},{"id":1215,"kind":"story","title":"Defense rocket motor startup Firehawk raises a Series D of $150 million","url":"https://wowtale.net/2026/10/31/267269/?utm_source=rss&utm_medium=rss&utm_campaign=%25eb%25a1%259c%25ec%25bc%2593%25eb%25aa%25a8%25ed%2584%25b0-%25ec%2596%2591%25ec%2582%25b0-%25ed%258c%258c%25ec%259d%25b4%25ec%2596%25b4%25ed%2598%25b8%25ed%2581%25ac-1%25ec%2596%25b55%25ec%25b2%259c%25eb%25a7%258c-%25eb%258b%25ac%25eb%259f%25ac-%25ec%258b%259c%25eb%25a6%25ac%25ec%25a6%2588d-%25ed%2588%25ac%25ec%259e%2590","host":"wowtale.net","by":"startupxo","score":0,"comments":0,"time":"2026-10-31T18:23:45.881Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["Firehawk Defense makes solid rocket motors that propel shells and missiles, and the propellant that goes into them. Solid propellant is a mass in which fuel and oxidizer are hardened together.","The conventional method mixes batches in pots and takes weeks to cure. The company claims that a continuous process, which melts and mixes thermoplastic materials with heat, produces the raw material within hours. Foreign media reported that a process that once took around 60 days was cut to the order of 6 hours.","The company stated that it aims to become the largest shell motor maker in the United States and the No. 3 propellant producer by the end of 2027. The company announced that Hydra-70 rocket live-fire tests at the Yuma test site in August were 100% successful.","Advent International and Rochefort Asset Management jointly led a Series D of at least $150 million. According to the company, its valuation is at least $1.4 billion, compared with the $290 million estimated by foreign media during the previous Series C round."],"discussion":"https://startupxo.com/en/1215/","markdown":"https://startupxo.com/en/1215.md","api":"https://startupxo.com/api/v1/item.json?id=1215&lang=en"},{"id":1214,"kind":"story","title":"How to connect to a private DB with AWS Session Manager without opening inbound ports","url":"https://qiita.com/j-yasukawa/items/701554f14fe8b02ff34e?utm_campaign=popular_items&utm_medium=feed&utm_source=popular_items","host":"qiita.com","by":"startupxo","score":0,"comments":0,"time":"2026-10-31T12:26:21.555Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["When a database sits on a network that cannot be reached directly from the internet, the traditional approach was to place an SSH relay server with port 22 open to the internet and manage key files. This article outlines how to connect to RDS or Aurora from a local DB client using AWS Systems Manager Session Manager without opening ports.","Session Manager works because the SSM Agent installed on the EC2 server opens an outbound HTTPS connection to AWS, so the security group needs no inbound rules. However, the author explains that the Agent cannot be installed on RDS, so one EC2 relay server is still required.","The configuration the author operates places Aurora PostgreSQL in a private subnet with no internet path, and keeps a t4g.nano relay server with no key pair and an empty inbound rule set turned off most of the time. The source presents this as the author's operational case and does not include cost or performance 3 verification results.","According to the source, access is controlled by the ssm:StartSession permission in IAM policies, and the start and end of sessions are recorded in CloudTrail. The source states explicitly that the SQL content passing to the database in port forwarding sessions is not recorded."],"discussion":"https://startupxo.com/en/1214/","markdown":"https://startupxo.com/en/1214.md","api":"https://startupxo.com/api/v1/item.json?id=1214&lang=en"},{"id":1213,"kind":"story","title":"Some Apple Intelligence features are coming to iPhones in China in the iOS 27.3 beta","url":"https://www.theverge.com/tech/1023974/apple-intelligence-china-ios-27-3","host":"theverge.com","by":"startupxo","score":0,"comments":0,"time":"2026-10-31T06:29:23.423Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["iPhone users in China will be able to use features such as notification summaries, live translation, writing refinement and proofreading, and removing objects from photos. The source presents these features as early features of Apple Intelligence, Apple's bundle of AI features.","In an announcement on Apple's China newsroom, Apple said it will release the iOS 27.3 beta on November 4 and bring the official version in December. This schedule was reported by MacRumors.","The latest Siri AI is not included in this China launch. Apple said it will connect Alibaba's AI chatbot Qwen for some experiences, and the source says this connection is similar to the ChatGPT integration.","Approval in China came in July, and Reuters reported that in August Apple separately trained an LLM for China. The source treats this approval as a sign that the launch is getting close."],"discussion":"https://startupxo.com/en/1213/","markdown":"https://startupxo.com/en/1213.md","api":"https://startupxo.com/api/v1/item.json?id=1213&lang=en"},{"id":1212,"kind":"story","title":"How XSS attacks slip malicious scripts into user input, and basic defenses","url":"https://qiita.com/Pigeon_gate/items/3ce4e1c1a39db64c5c32?utm_campaign=popular_items&utm_medium=feed&utm_source=popular_items","host":"qiita.com","by":"startupxo","score":0,"comments":0,"time":"2026-10-30T18:25:57.381Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["XSS is an attack that runs malicious scripts made by an attacker in other users' browsers. A script is a small program that makes a web page move. The source explains that just opening a malicious link or viewing a compromised page can make it run with the site's permissions.","The source divides XSS into reflected, stored, and DOM-based types. Reflected and stored XSS happen when the server builds dangerous HTML and sends it back. DOM-based XSS happens when the server sends only normal HTML and the browser's script inserts a URL value into the page, so server-side escaping is hard to rely on to block it.","The basic defense is to insert user input with textContent instead of innerHTML, so it is displayed only as text. The source also introduces Sanitizer APIs such as setHTML(). Citing web.dev material, it says Firefox supports them from 148 (2026 February), Chrome and Edge from 146, and Safari does not support them.","HttpOnly cookies stop scripts from reading cookies, which reduces the risk of session information theft. However, they do not stop script execution itself. CSP restricts execution of scripts that are not allowed, but the source says it has trade-offs: setting up external service integrations is troublesome and pages can easily break."],"discussion":"https://startupxo.com/en/1212/","markdown":"https://startupxo.com/en/1212.md","api":"https://startupxo.com/api/v1/item.json?id=1212&lang=en"},{"id":1211,"kind":"story","title":"Waymo to supply robotaxis based on Toyota electric vehicle bZ4X in Japan and the UK","url":"https://techcrunch.com/2026/10/30/waymos-next-robotaxi-will-be-a-toyota-bz4x-destined-for-japan-and-the-uk/","host":"techcrunch.com","by":"startupxo","score":0,"comments":0,"time":"2026-10-30T12:34:23.996Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["Toyota plans to produce the electric vehicle bZ4X 2027 model year, equipped with Waymo's autonomous driving system, at its Takaoka plant in Japan. Waymo plans to start test operations in Japan next year with this vehicle. The source adds that this vehicle is called BZ in North America.","Waymo has been testing Jaguar I-Pace self-driving cars in London, UK for several months and plans to start a robotaxi service once it receives the necessary permits. The source says Waymo also intends to deploy bZ4X in the UK and other right-hand-drive countries.","The two companies agreed to review a plan to apply Waymo's autonomous driving technology to privately owned Toyota vehicles, and a road safety project using Waymo's real operating data. The source says the specific details of the safety project are unclear."],"discussion":"https://startupxo.com/en/1211/","markdown":"https://startupxo.com/en/1211.md","api":"https://startupxo.com/api/v1/item.json?id=1211&lang=en"},{"id":1210,"kind":"story","title":"Can People Be Exempted from Using AI at Work on Religious Grounds? A Non-Religious Humanist Case","url":"https://techcrunch.com/2026/10/29/you-dont-need-to-believe-in-a-god-to-get-a-religious-exemption-from-using-ai-at-work/","host":"techcrunch.com","by":"startupxo","score":0,"comments":0,"time":"2026-10-30T06:37:27.185Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["David Dermen, an administrative staff member at the University of Kentucky, is a humanist who does not believe in God, but he said he received a religious accommodation (a work adjustment based on religious belief) that allowed him not to use AI at work. The university explained that it reviewed each request individually.","Dermen argued that AI threatens livelihoods, increases environmental burdens, and takes away human creative work. This is his claim, and the article has not confirmed it as fact.","A lawyer for the American Humanist Association (AHA) said that concerns about AI and human dignity can arise from many traditions, including Catholicism, Universalism, and humanism, and that people who do not believe in God also receive legal protection at work. A university spokesperson said the accommodation was provided on the condition that the employee continues to perform the core functions of the job."],"discussion":"https://startupxo.com/en/1210/","markdown":"https://startupxo.com/en/1210.md","api":"https://startupxo.com/api/v1/item.json?id=1210&lang=en"},{"id":1209,"kind":"story","title":"Impeccable, a skill that reduces AI design taste, compared across 4 ways with the same request","url":"https://qiita.com/jinxin4869/items/0e72c0b925b75f1c28c9?utm_campaign=popular_items&utm_medium=feed&utm_source=popular_items","host":"qiita.com","by":"startupxo","score":0,"comments":0,"time":"2026-10-30T00:50:25.381Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["Impeccable is a publicly available AI design assistant skill in the pbakaus/impeccable repository on GitHub. According to the original article, when installed in an AI tool such as Claude Code, it first asks for product information and sets a work order, which keeps the AI from falling back to a common average design.","The author made the same landing page request in four ways. The number of automated check findings was 5 to 8 items, so the difference was small. On the author's judgment of whether the design still works after changing the product name, only the method that asked for product information first and then built the page was hard to reuse for other products.","Run times were about 5 minutes for default Claude Code, about 15 minutes for the method that had existing results critiqued and revised, and about 20 minutes for the method that built from scratch with Impeccable. The author recorded these with a script in a Windows and WSL 2 environment with Claude Code v2.1.287, so these figures are values measured in that environment."],"discussion":"https://startupxo.com/en/1209/","markdown":"https://startupxo.com/en/1209.md","api":"https://startupxo.com/api/v1/item.json?id=1209&lang=en"},{"id":1208,"kind":"story","title":"Why are passkeys strong against fake sites? A summary of registration and login flows","url":"https://qiita.com/kuma_3838/items/1781531150821e2f0491?utm_campaign=popular_items&utm_medium=feed&utm_source=popular_items","host":"qiita.com","by":"startupxo","score":0,"comments":0,"time":"2026-10-29T12:34:48.025Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["A passkey is a login method that verifies you with a key stored on a device such as your phone or PC instead of a password. The original article explains that the browser mechanically compares the site address, and if it does not match, it does not show the key as a candidate, so login does not happen on a fake site.","The server stores only the public key, and the private key never leaves the device. The original article stresses that fingerprint or face data is also not sent to the server and is verified on the device.","The original article cites the Passkey Index released by the FIDO Alliance in 2025 October, saying that in a survey of 9 companies, the login success rate was 93% for passkeys and 63% for other methods, and the login time over 1 logins was 8.5 seconds versus 31.2 seconds. These figures were compiled by the FIDO Alliance and the research firm Liminal from numbers reported by participating companies.","The weakness of passkeys is the account recovery process. If attackers exploit the recovery path, the strength of passkeys can be reduced, so the original article introduces measures such as registering two or more passkeys and sending an alert to the account owner during recovery."],"discussion":"https://startupxo.com/en/1208/","markdown":"https://startupxo.com/en/1208.md","api":"https://startupxo.com/api/v1/item.json?id=1208&lang=en"},{"id":1207,"kind":"story","title":"Case study: Evaluating Japanese product search results with OpenSearch's LLM scoring feature and comparing them with human labels","url":"https://zenn.dev/nextbeat/articles/c22b45b1d784b0","host":"zenn.dev","by":"startupxo","score":0,"comments":0,"time":"2026-10-29T06:36:48.101Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["The author tested a Japanese product search by configuring OpenSearch's feature that scores how well search results match the query so that an LLM, instead of humans, does the scoring. The author calls this approach LLM_JUDGMENT.","The experiment used the Japanese portion of the Amazon product search dataset (ESCI). LLM scores were compared with human labels: E (exact match), S (substitute), C (complement), and I (irrelevant). The model was Gemini 3.5 Flash-Lite, and the environment was OpenSearch 3.9.0.","For the search 'speaker without power', a wooden unpowered speaker was scored 1.0, and USB-powered speakers were scored 0.2~0.3. A liquid formula milk was scored 1.0, the same as the human label E, but for the speaker cable, where the human label was E, the LLM gave 0.1.","The author attributed the gap to the LLM interpreting USB-powered and rechargeable speakers as 'has power', and noted that the cable case might also reflect variation in the human labels. Because this was an experiment with a few queries run by the author, it cannot be taken as a result that guarantees overall accuracy."],"discussion":"https://startupxo.com/en/1207/","markdown":"https://startupxo.com/en/1207.md","api":"https://startupxo.com/api/v1/item.json?id=1207&lang=en"},{"id":1206,"kind":"story","title":"Instagram Starts Requiring AI Accounts to Disclose AI Use, Mosseri Expects Backlash to Grow","url":"https://www.theverge.com/ai-artificial-intelligence/1021644/adam-mosseri-instagram-ai-negativity","host":"theverge.com","by":"startupxo","score":0,"comments":0,"time":"2026-10-29T00:49:22.151Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["Instagram head Adam Mosseri said in a podcast interview that anxiety and anger about AI are growing, especially in creative industries, and will likely grow further over the next few years. This is Mosseri's personal outlook.","Instagram said it has begun requiring AI accounts to label themselves. According to the source, it launched an AI account label feature earlier this year, and cracked down on AI-generated profiles in August.","Mosseri claimed Instagram does not rank AI creators differently. However, he said it is not certain whether AI content is rated as well as regular content.","AI content such as deepfakes that deceive celebrities remains a problem. Mosseri said that a full ban on AI content on platforms is not realistic."],"discussion":"https://startupxo.com/en/1206/","markdown":"https://startupxo.com/en/1206.md","api":"https://startupxo.com/api/v1/item.json?id=1206&lang=en"},{"id":1205,"kind":"story","title":"MacBook Pro with a touchscreen goes on sale November 6","url":"https://techcrunch.com/2026/10/28/apple-launches-its-new-touchscreen-macbook-pro/","host":"techcrunch.com","by":"startupxo","score":0,"comments":0,"time":"2026-10-28T18:29:40.299Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["A MacBook Pro whose screen you can press and control directly with your fingers has been released. The source reports that about 10 years after Apple added and then removed the Touch Bar, this is the product that fully supports touch input on a laptop screen.","The 14-inch model starts at 2,899 dollars, and the 16-inch model starts at 3,399 dollars. According to the source, pre-orders open October 30, and sales start November 6.","The chip can be M5 Pro or M5 Max. According to the source, memory goes up to 128GB, storage up to 16TB, and maximum display brightness is 2000 nits (HDR and SDR).","The source says Apple's designed hinge steadies the body when the screen is touched and also reduces camera shake during video calls. This is Apple's design explanation as relayed by the source and has not been independently verified."],"discussion":"https://startupxo.com/en/1205/","markdown":"https://startupxo.com/en/1205.md","api":"https://startupxo.com/api/v1/item.json?id=1205&lang=en"},{"id":1204,"kind":"story","title":"Review of taking the official Claude certification 4 exam, and notes on the Japanese exam and the December content revision","url":"https://qiita.com/minorun365/items/59783841e48f5467594f?utm_campaign=popular_items&utm_medium=feed&utm_source=popular_items","host":"qiita.com","by":"startupxo","score":0,"comments":0,"time":"2026-10-28T06:37:24.366Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["The Claude certification run by Anthropic is open to staff of companies that have joined the Claude Partner Network (a group of companies that work with Anthropic). According to the source, the exam became available in Japanese from the middle of October.","The author sat 4 exams over 2 weeks and passed 3 of them. The author wrote that CCDV-F and CCAR-P were scored at 90% or higher, while CCAR-F scored 702 points, which was 18 points short of the passing score of 720.","The source cites Pearson VUE guidance saying that the content of CCAO-F, CCDV-F, and CCAR-P changes from December 7. The author recommended taking the exam in November, before this revision, to feel more secure.","Official materials are mostly videos and web content and are hard to read through, so the author had Claude organize the study material, created a 4-choice practice question set, and solved it on a smartphone. The author wrote that CCAR-P was passed after solving this question set 1 times."],"discussion":"https://startupxo.com/en/1204/","markdown":"https://startupxo.com/en/1204.md","api":"https://startupxo.com/api/v1/item.json?id=1204&lang=en"},{"id":1203,"kind":"story","title":"Google DeepMind's world model Genie 4 opens first to trusted testers as a real-time simulation that changes scenes with words","url":"https://deepmind.google/blog/genie-4-a-new-generation-of-world-simulation/","host":"deepmind.google","by":"startupxo","score":0,"comments":0,"time":"2026-10-28T00:50:02.932Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["Genie 4 is a world model that creates a virtual world from a text description of a scene, which users can explore by moving through it with a keyboard. The source says instructions to change the weather or add people and objects are also reflected in real time.","According to the source, Genie 4 generates sound together with video, including ambient sounds, music, and character dialogue. The company said this is the first time a real-time world model has generated sound directly.","The source says Gemini 3.8 Flash watches the generation process, divides goals into several steps, and checks whether each step succeeded. This is the design explanation offered by Google, and the source does not include external verification results.","The release starts as a research preview for trusted testers. Google AI Ultra subscribers will first have access through Project Genie, and early partners will have access through the Gemini API and Google AI Studio. The source does not specify when general release will happen."],"discussion":"https://startupxo.com/en/1203/","markdown":"https://startupxo.com/en/1203.md","api":"https://startupxo.com/api/v1/item.json?id=1203&lang=en"},{"id":1202,"kind":"story","title":"Recording Claude Code session notes makes design reasons and work history easier to find later","url":"https://qiita.com/qwertyhoge/items/987c0ea8afdb5cd045e8?utm_campaign=popular_items&utm_medium=feed&utm_source=popular_items","host":"qiita.com","by":"startupxo","score":0,"comments":0,"time":"2026-10-27T18:28:54.805Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["Developer @qwertyhoge introduces a self-made work instruction file (skill) that they call with the /decision-log command when a Claude Code task ends. It selects technical choices, alternatives not adopted and their reasons, parts where AI suggestions were corrected, whether self-review was done before the PR, and whether code license and confidential information were checked, then appends them to a monthly record file.","When there is an important choice, they also write an ADR (Architecture Decision Record, a document that records the background and alternatives of a design decision). According to the original, over about 1 months they left 5 work records and 10 ADRs.","The author explains that AI tends to interpret its own work favorably, so they added an instruction not to write about things that were not done. They wrote that across several tests with the Opus 5.5 model they did not see cases of it writing about things that did not happen, but this is the author's personal observation.","Looking at the records, the author found they had rarely done self-review before PRs. They therefore created a separate skill and hook that enforce review, so a PR cannot be created without going through review."],"discussion":"https://startupxo.com/en/1202/","markdown":"https://startupxo.com/en/1202.md","api":"https://startupxo.com/api/v1/item.json?id=1202&lang=en"},{"id":1201,"kind":"story","title":"Google Workspace Weekly News: Sheets Box Plot and Mobile Voice Document Creation Added","url":"https://workspaceupdates.googleblog.com/2026/10/weekly-recap-10-23-2026.html","host":"workspaceupdates.googleblog.com","by":"startupxo","score":0,"comments":0,"time":"2026-10-27T15:15:51.646Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["Google Sheets now includes a box plot chart as a standard feature, showing how numbers are spread out. When you select a range of numbers, it calculates and displays the minimum, median, maximum, and outliers without a separate formula.","Docs Live in the Google Docs mobile app (Android, iOS) is a feature that creates structured documents through conversational speech. According to the original text, it will roll out in stages to eligible Workspace users.","Real-time speech translation in Google Meet has been expanded to support 63 languages along with automatic language detection. The original text says it is available on web and mobile (Android, iOS), and that on Meet-dedicated devices it is listen-only.","Google Vault can now export Gemini conversations from Workspace filtered by conditions. Admins can target conversations from Gmail, Drive, Docs, Sheets, Slides, Meet, and others, and the original text says Vault's security and privacy controls apply as they are."],"discussion":"https://startupxo.com/en/1201/","markdown":"https://startupxo.com/en/1201.md","api":"https://startupxo.com/api/v1/item.json?id=1201&lang=en"},{"id":1200,"kind":"story","title":"Autonomous kit that attaches to existing excavators expands to real construction sites","url":"https://techcrunch.com/2026/10/26/this-startups-self-driving-excavators-help-dig-up-600m-in-new-funding/","host":"techcrunch.com","by":"startupxo","score":0,"comments":0,"time":"2026-10-26T18:48:19.440Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["Bedrock Robotics makes an autonomous driving kit that attaches to existing equipment such as excavators at construction sites. According to the source, the kit can be installed on construction vehicles and other work vehicles through retrofitting.","According to the company, its autonomous excavators are in commercial operation at a Nevada water treatment facility, a Texas civil site, and 1.2 million cubic-yard-scale civil works. This is based on the company's announcement, and the source does not include data on work performance verified by a 3 party.","The company claims it takes more than 5 years to become skilled at operating an excavator. The source cites an estimate from the American trade group Associated Builders and Contractors that 40% or more of construction workers will retire within 5 years.","This Series C round (a funding round at the growth stage) is $600 million in size, and the company valuation was reported as $4.1 billion. The company said it plans to use the funds to deploy its kit at more sites and to expand the staff and infrastructure that support customers."],"discussion":"https://startupxo.com/en/1200/","markdown":"https://startupxo.com/en/1200.md","api":"https://startupxo.com/api/v1/item.json?id=1200&lang=en"},{"id":1199,"kind":"story","title":"An instruction file that first explains AI-written technical text in plain words, then shortens it","url":"https://zenn.dev/microsoft/articles/explain-before-shorten","host":"zenn.dev","by":"startupxo","score":0,"comments":0,"time":"2026-10-26T14:10:20.721Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["A personal opinion post by Kazuki Ota, a Microsoft employee in Japan, published on Zenn in the 'Microsoft (有志)' publication. It covers the problem where AI-written technical explanations start with a negation such as '0건이라는 뜻은 아니다', which blurs the key point.","The author's method is to first decide in one sentence what the explanation should convey, explain it in plain words as if speaking to a colleague, and then shorten it while keeping the meaning. The author says this procedure was organized into an AI instruction file called SKILL.md.","As an example, 'separate acceptance and completion checks' is changed to 'notify acceptance first without waiting for processing to finish, and look up the result later'. The focus is on keeping the meaning intact when the wording is shortened.","The source contains no data measuring or comparing the effect of this method. The author subjectively evaluated the results as fairly good."],"discussion":"https://startupxo.com/en/1199/","markdown":"https://startupxo.com/en/1199.md","api":"https://startupxo.com/api/v1/item.json?id=1199&lang=en"},{"id":1198,"kind":"story","title":"Node.js Brotli decompression issue that treated malformed input as insufficient input, and the fix","url":"https://zenn.dev/chot/articles/79e6c94e27782d","host":"zenn.dev","by":"startupxo","score":0,"comments":0,"time":"2026-10-26T07:02:25.350Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["Node.js uses a built-in decoder when it decompresses Brotli-compressed data. When this decoder received malformed input, it could write outside its internal buffer. The issue was fixed on October 21, 2026 in a Node.js security release as CVE-2026-111497. The impact description (crash, memory exposure, possible remote code execution) is quoted from the Node.js vulnerability advisory.","The cause was that an error in a block switch (an instruction to change the compression table used for the next section) was handled as a return value such as 'more input needed'. According to the quoted advisory, this caused the malformed data to be treated as a temporary state, and input kept accumulating in the internal buffer.","On the unsafe execution path, where enough input was available, the same error was not passed to the caller and was ignored. The fix changed the return type to an error code enum and added the new error code BROTLI_DECODER_ERROR_FORMAT_BLOCK_SWITCH, so both paths now stop processing immediately when an error occurs.","The core of the fix is bringing the upstream google/brotli commit into Node.js. This article is an explanation written after reading the fix commit. The attack path and the actual scope of harm are not covered in the original text."],"discussion":"https://startupxo.com/en/1198/","markdown":"https://startupxo.com/en/1198.md","api":"https://startupxo.com/api/v1/item.json?id=1198&lang=en"},{"id":1197,"kind":"story","title":"iFixit's teardown of the iPhone Duo: how the foldable screen hides creases and how hard it is to repair","url":"https://www.theverge.com/tech/1018233/ifixit-iphone-duo-teardown-repairability-hinge","host":"theverge.com","by":"startupxo","score":0,"comments":0,"time":"2026-10-26T01:05:51.792Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["After taking apart the iPhone Duo, iFixit gave it a repairability score of 5 out of a possible 10. According to the source, this is the highest score the site has given to a foldable phone.","The matte nano-texture cover on the inner screen makes the display easier to read in bright sunlight and scatters reflections so the creases are less visible. iFixit estimated that the oval subpixel layout seen under a microscope is also designed to reduce creases.","The source explains that repairing the screen itself is expensive and difficult. The outer nano-texture layer can be replaced relatively easily without changing the OLED underneath.","On reports of broken screens and cracked hinges, iFixit offered a hypothesis: the hinge sometimes does not fully close, and the accordion cable running near the edge of the battery may concentrate pressure at that spot. Apple did not immediately respond to a request for comment."],"discussion":"https://startupxo.com/en/1197/","markdown":"https://startupxo.com/en/1197.md","api":"https://startupxo.com/api/v1/item.json?id=1197&lang=en"},{"id":1196,"kind":"story","title":"Feature guide: Using Copilot in OneDrive to gather multiple tables and notes into a report","url":"https://note.com/kaiyu06/n/n724f34b96950?hl=en","host":"note.com","by":"startupxo","score":0,"comments":0,"time":"2026-10-25T18:33:42.408Z","origin":"curated","language":{"source":"ko","reading":"en","translated":true},"summary":["Microsoft announced the general availability of Copilot in OneDrive and SharePoint. Copilot is the name of an AI feature offered by Microsoft.","The use cases presented were analysis of multiple tables, folder organization, and conversational report writing based on existing material. An example of saving a repetitive process as a personal skill was also presented.","The announcement includes features that will be rolled out in sequence over the coming months. The source states that not all of the introduced features are provided at once, and that the actual reduction in work time was not measured."],"discussion":"https://startupxo.com/en/1196/","markdown":"https://startupxo.com/en/1196.md","api":"https://startupxo.com/api/v1/item.json?id=1196&lang=en"}]}