# StartupXO: Newest (en), page 7

> Every live post, newest first. HTML: https://startupxo.com/en/new/?p=7

181. [Show XO: CharaSync — AI images and videos with reusable character references](https://charasync.com/) (charasync.com)
   1 point by terencexu003, 2026-10-20T12:11:02.595Z | 1 comment | https://startupxo.com/en/1148/ | https://startupxo.com/en/1148.md
182. [Some Google Home speakers keep downloading hundreds of GB of data per day after a failed update](https://www.theverge.com/tech/1014209/google-home-speaker-bug-continuous-download) (theverge.com)
   0 points by startupxo, 2026-10-20T10:28:20.842Z | 0 comments | https://startupxo.com/en/1147/ | https://startupxo.com/en/1147.md
   - Some Google Home speaker users are experiencing devices that keep downloading large amounts of data in the background, according to The Verge. The original article says a failed firmware update installation may be the cause, but does not treat this as a confirmed cause.
   - According to the original article, one Reddit user reported that the device uses more than 200GB per day. Comparing the same period, a home speaker in another household used 2GB over one month, while the affected device used 132GB, according to 9to5Google.
   - Google said a fix is being rolled out for 9to5Google. This is Google's explanation, and the original article does not confirm which devices the fix applies to.
   - Most reports focus on newer Google Home speaker models. However, the full scale of the problem is also unclear in the original article.
183. [DeepSeek open-sources DeepInfer, a large-scale AI model inference engine](https://github.com/deepseek-ai/DeepInfer) (github.com)
   0 points by startupxo, 2026-10-20T10:28:05.934Z | 0 comments | https://startupxo.com/en/1146/ | https://startupxo.com/en/1146.md
   - DeepInfer is software that speeds up the process by which an AI model receives a question and produces an answer. The source says DeepSeek uses it as the engine for its own large-scale production service.
   - This release focuses on offline inference. According to the source, the KV cache system, separation of prefill and decode, and load balancing across data-parallel replicas that a complete service needs are not included in this release.
   - The preparation environment listed in the source is Linux, GCC 13 or later, and Python 3.12. NVIDIA GPUs require CUDA 13, and Huawei Ascend NPUs require CANN 9.2. The source lists the supported hardware as NVIDIA SM100 GPUs and Ascend 950 NPUs.
   - The performance figures in the README say they were measured with 4 prefill SM100 GPUs and 64 decode GPUs within the same high-speed interconnect group. These are results reported by the development team that published the repository, and 3-party verification is not confirmed in the source.
184. [WhatsApp help page hidden text contains instructions telling the chatbot not to give regulator information](https://redact.dev/blog/whatsapp-help-pages-hidden-ai-instructions) (redact.dev)
   0 points by startupxo, 2026-10-20T08:38:21.764Z | 0 comments | https://startupxo.com/en/1145/ | https://startupxo.com/en/1145.md
   - Original author Redact saved and reviewed 545 WhatsApp help pages and said AI instructions were found in 102 hidden text blocks across 80 pages. The claim is that users could not see them.
   - The hidden instruction on the account deletion page told the chatbot not to suggest contacting regulators or using another phone to people who lost their phone and could not delete the account in the app. WhatsApp told WIRED the instruction was for its support chatbot and was exposed due to a bug.
   - The hidden instruction on the report and block page prohibited citing privacy laws such as GDPR, LGPD, and CCPA when users mentioned lawyers or lawsuits. The original article compares this with WhatsApp's European privacy policy, which publicly lists the same laws and the right to complain to supervisory authorities.
   - According to an update to the original article, when checked on 2026 October 18, all 102 hidden blocks had disappeared. However, the original article does not confirm whether the hidden instructions were actually used in real chatbot answers.
185. [Estonian startup Blackwall raises $100 million for website bot traffic management tools](https://techcrunch.com/2026/10/20/blackwall-quadruples-valuation-to-2-2b-as-websites-wrangle-an-increasingly-crowded-internet/) (techcrunch.com)
   0 points by startupxo, 2026-10-20T08:37:39.957Z | 0 comments | https://startupxo.com/en/1144/ | https://startupxo.com/en/1144.md
   - Estonian company Blackwall makes security tools that help hosting providers and websites tell apart bot traffic, especially visits from AI agents.
   - The company said it raised $100 million in a Series C round and that its valuation is $2.2 billion. Roy Karti, Blackwall's executive chairman, said the valuation has grown 4 times since its investment before 2025.
   - Blackwall does not sell directly to websites. It offers a bundle of tools to hosting, managed service, and cloud providers. These providers sell to their own customers under their own brands and share revenue with Blackwall.
   - Customers can use a free tier that shows traffic and threat status. Real protection features and tools require payment. The company said it uses digital fingerprinting and its own AI models to separate requests sent by people from malicious bots. This is the company's own explanation, and the source does not include independent verification results.
186. [Auto feature rolls out to all users, using past conversations to help start the next tasks from the dashboard](https://blog.autohive.com/whats-new-this-week-9-16-october-2026/) (blog.autohive.com)
   0 points by startupxo, 2026-10-20T06:43:55.512Z | 0 comments | https://startupxo.com/en/1143/ | https://startupxo.com/en/1143.md
   - Auto has become the default assistant (agent) on the workspace dashboard and is available to all users. The company said it can read past conversations and help with work by creating agents and skills (predefined ways of working).
   - New sign-ups who choose their industry and frequently used apps receive suggestions for tasks worth automating and their first agent setup. The original text includes no figures showing this feature's effect.
   - Instagram integration adds carousel posts that let users swipe through multiple images and posts that mix images and videos, and upload stability has also improved, the company said.
   - Gmail integration now cleans up the email body text that agents read, loads results in smaller units, and lets long emails and conversation threads be passed along in pieces. Browser notifications now appear as system notifications instead of on-screen pop-ups.
187. [An open-source tool that disables Android TV's built-in apps over Wi-Fi and brings up an ad-free home screen](https://github.com/Gathin23/cleanyourtv) (github.com)
   0 points by startupxo, 2026-10-20T04:33:17.530Z | 0 comments | https://startupxo.com/en/1142/ | https://startupxo.com/en/1142.md
   - This is a GitHub repository that explains how to turn off manufacturer apps and ad rows on Android TV and Google TV and replace an ad-heavy home screen with a lightweight launcher. The source says it disables apps rather than deleting them, so a restore script can undo the changes.
   - You need a computer on the same Wi-Fi network, a coding assistant that can run shell commands (for example, Claude Code), and USB debugging enabled on the TV. The creator wrote that following the repository takes about 15 minutes, and this is the creator's own estimate.
   - According to the source, it covers 66 TV models and 25 brands. The source says Samsung (Tizen) and LG (webOS) run different systems, so the same method cannot be used as is, and that on Fire TV only the app cleanup part works.
   - The creator says the launcher was built for TVs with 1GB RAM, and the 300KB size and lightness are also the creator's claims. The source does not include independent performance test results.
188. [OpenBSD 8.0 released with wireless roaming, laptop sleep resume and other improvements](https://www.openbsd.org/80.html) (openbsd.org)
   0 points by startupxo, 2026-10-20T04:32:31.383Z | 0 comments | https://startupxo.com/en/1141/ | https://startupxo.com/en/1141.md
   - The OpenBSD project published a release notice for 8.0. The notice says it was released on October 20, 2026, and is OpenBSD release number 61.
   - According to the original list, the qwx-family wireless LAN driver fixes roaming problems between access points, adds support for 40MHz channels, and adds support for 802.11n. These changes may change how wireless connections behave.
   - Several sleep and hibernate fixes are included. The original mentions a change so that laptops do not fail to wake after the lid is closed, and a change that checks disk image size when resuming from hibernate.
   - Many security fixes are also included. The original says vulnerabilities and memory errors were fixed in perl, libexpat, the pf firewall, the httpd web server, dhcpd and others, as stated in the project notice. The actual scope of impact is outside the original text and has not been confirmed.
189. [Baseus U2 Pro, security camera with a light that follows moving targets](https://www.theverge.com/tech/1005073/baseus-security-primesentinel-u2-pro-outdoor-3k-dual-camera-tracking-spotlight) (theverge.com)
   0 points by startupxo, 2026-10-20T04:32:10.219Z | 0 comments | https://startupxo.com/en/1140/ | https://startupxo.com/en/1140.md
   - The product mounts a light on a motor-driven mount that turns toward moving targets, so the owner quickly knows when a possible intruder has been detected. The source compares this with existing security cameras, which mainly record and send alerts.
   - According to the source, it uses one fixed 3K camera and one 3K camera that can zoom, tilt, and also rotate 360. Color night footage is said to make it easier to see the colors of people, animals, and vehicles.
   - AI analysis is said to tell apart people, pets, vehicles, and delivery boxes for alerts, and to ignore unrelated movement such as tree branches. The source states this is a feature the manufacturer Baseus has described, and the source contains no independent verification results.
   - Video is stored only when movement is detected, on a 512GB encrypted memory card. Remote viewing is done through the Baseus app, and live views can also be seen on Amazon Alexa or Google Home devices. The installation time of about 2 minutes and battery use of up to 150 days are manufacturer claims.
190. [Hugging Face policy head says AI evaluation methods are not yet mature and independent evaluators are difficult](https://techcrunch.com/2026/10/19/embedded-ai-evaluators-wont-be-an-easy-solution-hugging-faces-policy-chief-says/) (techcrunch.com)
   0 points by startupxo, 2026-10-20T02:37:20.315Z | 0 comments | https://startupxo.com/en/1139/ | https://startupxo.com/en/1139.md
   - Hugging Face's chief policy officer Irene Solaiman said at a TechCrunch Disrupt panel that AI model evaluation methods are still evolving.
   - Solaiman explained that to trust evaluation results, reproducibility, meaning whether the same results can be obtained again, matters, and whether open weights are released greatly affects this.
   - The target of evaluation is expanding from the model itself to the harness (the tool environment through which users access the model) and the whole agent. Solaiman said that differences in the harness can lead a weaker model to produce better results than a stronger model.
   - A method of placing a neutral 3rd-party evaluator inside AI companies is drawing attention. The source reports that after Anthropic hired the consulting firm Accenture as its first evaluator, there was criticism about its qualifications.
191. [Flipkart Tests Food Delivery Service with About 8,000 Employees](https://techcrunch.com/2026/10/19/walmarts-flipkart-is-quietly-building-a-food-delivery-challenger-in-india/) (techcrunch.com)
   0 points by startupxo, 2026-10-20T02:36:03.386Z | 0 comments | https://startupxo.com/en/1138/ | https://startupxo.com/en/1138.md
   - Flipkart, the e-commerce company owned by Walmart, is first testing its food delivery service with its employees. According to two anonymous sources, the number of testers has grown from several hundred at first to more than 8,000 people.
   - The test service is offered as a separate tab inside the shopping app. Restaurant listings come from ONDC (a network linking buyers and sellers across multiple platforms), a government-backed online trade network, and from restaurants that Flipkart contracted directly. Deliveries are handled by a dedicated fleet of riders from Ekart, its logistics subsidiary.
   - Sagar Dariani, head of a restaurant association, said Flipkart told restaurants it plans a full launch at the end of October and expansion to 5~6 cities by year-end. The company did not respond to a request for comment.
   - Dariani said the base commission Flipkart offered is about 10%, lower than the 20~25% charged by existing providers. He also claims that, once advertising and discounts are added, existing platforms cost 40~50% of order value.
192. [Final Porta over-ear wireless headphones weighing 40 grams, a lightweight option to wear on your ears](https://www.theverge.com/tech/1013635/its-hard-to-imagine-wireless-headphones-ever-getting-lighter-than-these) (theverge.com)
   0 points by startupxo, 2026-10-19T18:38:51.504Z | 0 comments | https://startupxo.com/en/1137/ | https://startupxo.com/en/1137.md
   - Final's new Porta is an over-ear wireless headphone that rests on the ears. The source says the headband uses a single wire structure, which reduces the weight to 40 grams.
   - According to the source, the Marshall Milton A.N.C., in the same over-ear category, weighs 200 grams. It may be useful for comparison by users who find earbuds or ear-hook types uncomfortable.
   - In Japan, it can be chosen with black or orange ear pads for 12,800, about 80 according to the source, and it is open for pre-order at online stores. The source says shipping is scheduled to begin on October 23.
   - The battery is 57mAh, it charges via USB-C, and the source says it offers up to 8 hours of use. It supports Bluetooth 5.4 with SBC and AAC codecs and has no noise cancelling. The source explains that the IPX4 water resistance rating means it is fine when exposed to rain or sweat.
193. [Open-source project Mac.apk runs Android apps directly on Mac without an emulator](https://github.com/kksimp/Mac.apk) (github.com)
   0 points by startupxo, 2026-10-19T18:37:24.500Z | 0 comments | https://startupxo.com/en/1136/ | https://startupxo.com/en/1136.md
   - Mac.apk is a project that runs Android app installation files (APK) directly on Apple silicon Macs. The source says that apps run like regular Mac programs without starting a virtual machine, an emulator, or Linux.
   - According to the source, apps appear in the Dock, have their own window with a back button, can be resized like Mac windows, and follow light and dark mode. The claim that apps run at native speed comes from the development team, and the source contains no independent measurement results.
   - The requirements are macOS 26.4 (Tahoe) or later and an Apple silicon Mac. The source states that Intel Macs are not supported. The program itself is about 280MB, and each installed app takes up roughly 1.5 to 3 times the APK size in storage.
   - The project is labeled as alpha. The source says some apps work well, while others work only partially or do not work yet. Test results for each app can be checked in the compatibility list, and the source code is released under the AGPL-3.0 license.
194. [Apple previews October 28 product event, with new Macs and an iPad mini expected](https://www.theverge.com/tech/1013028/apple-second-october-product-launch-announcement) (theverge.com)
   0 points by startupxo, 2026-10-19T18:36:48.047Z | 0 comments | https://startupxo.com/en/1135/ | https://startupxo.com/en/1135.md
   - According to a report by The Verge reporter Jay Peters, Apple has sent invitations to a product event on October 28 at 1 PM (Eastern US time, Los Angeles local time). Marketing chief Greg Joswiak also posted on X previewing next week's announcements.
   - According to the original report, which cites Bloomberg, Apple plans to unveil new Macs and an improved iPad mini at this event. The biggest announcement is expected to be a new MacBook Pro with an OLED touchscreen, and reports say a new 14-inch MacBook Pro and a new iMac using the M6 chip will also be shown.
   - The original report says the new MacBook Pro will be lighter than previous models and will have a function area similar to the Dynamic Island at the top of the iPhone screen. This is an expectation based on the report, and Apple has not confirmed it.
   - The iPad mini is reported to use an OLED screen, with the selfie camera moved to the landscape edge. Earlier, at a smart-home-focused event on October 13, Apple unveiled the new HomeView home hub, an updated Apple TV 4K, and the 2-generation HomePod mini.
195. [Amazon agrees to fund nuclear power plant construction in Indiana](https://www.theverge.com/tech/1013735/amazon-x-energy-aep-nuclear-indiana) (theverge.com)
   0 points by startupxo, 2026-10-19T16:34:28.006Z | 0 comments | https://startupxo.com/en/1134/ | https://startupxo.com/en/1134.md
   - This plant aims to supply nuclear-generated electricity to both the local power grid and Amazon data centers. The source reports that it plans up to 960 megawatts of carbon-free power.
   - Amazon agreed with X-energy and the US utility AEP to fund construction of the plant in Fairbanks, Indiana. The plan would replace the retired Breed coal plant site.
   - The reactors are planned to consist of 12 Xe-100 small modular reactors. The source explains that these reactors are still under development and licensing and are a technology that has not been fully verified.
   - Amazon said that up to 1,000 jobs could be created during construction and 100 permanent jobs after operation begins. This is Amazon's projection, and the construction schedule and contract terms have not been disclosed.
196. [Google Play adds on-device running of personal AI-built apps and a taste-based game finder](https://blog.google/products-and-platforms/platforms/google-play/google-play-updates-october-2026/?authuser=91&hl=bn) (blog.google)
   0 points by startupxo, 2026-10-19T16:33:41.019Z | 0 comments | https://startupxo.com/en/1133/ | https://startupxo.com/en/1133.md
   - Android apps built with Google AI Studio can now run on your own device without a developer account. According to the source, this will roll out in phases starting this day in the US and some other markets, for users aged 18 or older.
   - According to the source, personal apps stay private and separate from public store listings. A feature to share with friends only is planned for later. To publish to the world, the app must switch to a professional developer account and go through the same review as regular developers.
   - Game Finder recommends games based on Gamer DNA, which analyzes preferences and play styles that users choose to share. According to the source, it starts today for English-speaking users in the US, and other markets will follow later.
   - According to the original announcement, 48-hour short Sprints competitions are now available in all markets, and community posts where players share questions and answers on game detail pages have also opened. This is Google Play's own announcement, and the source contains no independent verification of its effects.
197. [More users are trying to strip ads and unnecessary apps from smart TVs with AI](https://techcrunch.com/2026/10/19/people-are-using-ai-to-make-their-smart-tvs-dumb-again/) (techcrunch.com)
   0 points by startupxo, 2026-10-19T16:33:19.202Z | 0 comments | https://startupxo.com/en/1132/ | https://startupxo.com/en/1132.md
   - Methods for reducing ads, slow menu screens, and viewing tracking on smart TVs by instructing AI coding tools are being shared. The source calls this vibe coding, which means creating code by giving instructions in natural language.
   - The dumbtv project published on GitHub aims to block ads and viewing tracking built into the TV itself. The project side says it tested on TVs from several manufacturers, including Fire TV, but this is a claim by the development team.
   - cleanyourtv removes unnecessary apps from Android TV and Google TV over Wi-Fi and replaces the home screen with one similar to Apple TV. Its creator wrote that they removed 25 Xiaomi-related packages from their parents' TV.
   - According to the source, one user claimed that by giving only the TV model information to AI, they removed apps and sped up the menu, and another user shared results from disabling apps and changing the standby screen. All are individual reports posted by people and have not been independently verified.
198. [Small model further trained to reduce AI-style writing, can run on a laptop](https://100toolkit.com/blog/deslop-model/) (100toolkit.com)
   0 points by startupxo, 2026-10-19T14:32:02.239Z | 0 comments | https://startupxo.com/en/1131/ | https://startupxo.com/en/1131.md
   - Qwen3-4B model was further trained to reduce writing habits that look AI-generated. According to the creator, it uses fewer dashes, fewer 'not X but Y' sentences, and fewer lists grouped in threes, and it can be used without a separate system prompt.
   - The creator's own rule-based detector scored it at 1000 per word, dropping from 121 for the original model to 12.5. This score is based on the creator's regex criteria, and no blind test with readers has been done yet.
   - The creator said the training data included 161 answers generated on a MacBook Air. The creator also reported that it got all 30 answers right in a quiz of 30 factual questions, and this is the creator's own measurement.
   - The original post says it can be run in Ollama by downloading the Q4_K_M quantized file (about 2.5GB). The original weights and the LoRA adapter are also published.
199. [Wyze's new scale moves the display to the handle, making readings easier at eye level](https://www.theverge.com/tech/1013081/wyze-smart-scale-ultra-bodyscan-plus-handle-display) (theverge.com)
   0 points by startupxo, 2026-10-19T14:31:29.591Z | 0 comments | https://startupxo.com/en/1130/ | https://startupxo.com/en/1130.md
   - Wyze unveiled the Ultra BodyScan Plus, which moves a 4.3-inch color display onto a foldable handle to show measurements close to eye level, according to the source.
   - The price is 149, about 30 more than the previous Ultra BodyScan (119.98). According to the source, the earlier model measured 13 items, including weight, BMI, body water and muscle mass, and this one expands that to 50 or more, including visceral fat, protein and basal metabolic rate.
   - It provides analysis of arms, legs and torso using BIA (bioelectrical impedance analysis, a method that passes a weak current through the body to estimate body composition). The source says it uses an ITO (indium tin oxide) coated glass surface so measurements work regardless of foot position. These figures and accuracy claims are Wyze's own, and the source contains no independent verification.
   - According to the source, data can be sent to Apple Health, Google Health and Android Health Connect without a subscription, either by Bluetooth through a smartphone or directly to the cloud over home Wi-Fi. The source says competitor Withings requires new users to pay an annual subscription of 99.50.
   - Up to 8 users can be registered, and guests can take one-time measurements without limit. There is also a mode for weighing children, pets or yourself while holding items, but detailed health metrics are not provided.
200. [Apple tightens guidelines to ban apps that hide nude-synthesis features until after App Review approval](https://techcrunch.com/2026/10/19/apple-updates-guidelines-to-ban-nudify-apps-masquerading-as-regular-apps-to-pass-review/) (techcrunch.com)
   0 points by startupxo, 2026-10-19T14:31:01.711Z | 0 comments | https://startupxo.com/en/1129/ | https://startupxo.com/en/1129.md
   - The original article reports that Apple revised its App Review guidelines to ban apps that use AI to create offensive content. Nude-synthesis apps are apps that create images of people in photos with their clothes removed.
   - According to the original article, Apple bans apps that appear to be ordinary apps during review and then reveal nude-synthesis features after approval. This tactic is called bait-and-switch. The policy also says apps that add features after submission that were not present at submission will be removed.
   - An Apple blog post says apps that handle user-uploaded content must filter out offensive material and provide a reporting function and a way to block abusive accounts. The original article does not confirm how strictly this will be enforced.
   - According to the original article, Apple blocked more than 150,000 apps from release during review last year and removed 190,000 3thousand developer accounts over fraud and abuse. The city of San Francisco ordered Apple and Google to remove nude-synthesis apps in July.
201. [Mistral releases Timestral, a model that forecasts future values from past numbers](https://mistral.ai/news/timestral/) (mistral.ai)
   0 points by startupxo, 2026-10-19T12:43:10.577Z | 0 comments | https://startupxo.com/en/1128/ | https://startupxo.com/en/1128.md
   - Timestral is a family of models that looks at numbers recorded in time order, such as equipment sensor readings or daily sales, and predicts future values. It comes in four sizes, from 5M to 209M parameters, and instead of a single predicted value it shows a range of uncertainty using several quantiles.
   - Mistral says its largest 209M model finished 1 place on three public benchmarks: GIFT-Eval, TIME, and fev-bench. This is the developer's claim. The comparison was made among zero-shot models that are used without additional training, using a single checkpoint for each.
   - According to the source, on a 8-thread server CPU, forecasting one day ahead from 3 series recorded at 15-minute intervals took 10ms for the 5M model and 140ms for the 209M model. These speed figures are also conditions measured and reported by Mistral, and the source explains that inference can run without a GPU.
202. [Maeving RM2 electric motorcycle: style and quiet ride are good, but finish and price disappoint](https://www.theverge.com/transportation/1013049/maeving-rm2-electric-motorcycle-review-specs) (theverge.com)
   0 points by startupxo, 2026-10-19T12:42:43.279Z | 0 comments | https://startupxo.com/en/1127/ | https://startupxo.com/en/1127.md
   - This motorcycle runs on an electric motor inside the rear wheel instead of an engine and chain. Reviewer Tim Stevens named its quietness, with almost no drivetrain noise, as its biggest advantage.
   - The rated top speed is 70mph. The fastest figure the reviewer recorded in the strongest mode was about 63mph, leaning forward pushed it up to 67mph, and 70mph was reached only downhill. The measurement was based on the reviewer's own riding feel.
   - The official range is 80 miles per charge. The reviewer thought the figure was only likely in the default mode (limited to 45mph), and wrote that in sport mode the average was about 50 miles per charge.
   - The battery is a removable 2.73 kilowatt-hour (kWh) pack, and there are 2 packs. Each pack weighs 36 pounds. The bike's charging port uses a proprietary standard, so you must buy a standard J1772 adapter separately (440 dollars), and this adapter does not increase charging speed. Charging 80% with the basic charger took about 4 hours.
203. [More people are using AI coding tools to fix firmware on music gear themselves](https://www.theverge.com/gadgets/1013493/vibe-coding-ai-music-gear-hacking-custom-firmware) (theverge.com)
   0 points by startupxo, 2026-10-19T12:42:02.108Z | 0 comments | https://startupxo.com/en/1126/ | https://startupxo.com/en/1126.md
   - Vibe coding is a way of describing the function you want in words and having AI generate the code. According to the article, this method has greatly increased the number of hobby developers who modify firmware, the internal programs of instruments such as synthesizers.
   - The article's reporter observed that adding a single velocity-sensitive feature to the Korg Volca used to take weeks to months, but now hobby developers turn low-cost devices into grooveboxes in a few days.
   - Several custom firmwares have appeared for the M-Vave FM-1, a DX-7 compatible synthesizer priced at about 70. The article reports that some firmware pages on GitHub list Claude as a contributor or disclose AI use, while Baudgirl only says AI is suspected to have been used.
   - Elektron warned on its forum that it bans links to custom firmware and that devices can be bricked. By contrast, M-Vave has released the open-source firmware FM1 Open and a development kit.
204. [How the overhead projector went from military technology to a classroom tool](https://www.theverge.com/podcast/1013473/overhead-projector-version-history) (theverge.com)
   0 points by startupxo, 2026-10-19T10:32:03.653Z | 0 comments | https://startupxo.com/en/1125/ | https://startupxo.com/en/1125.md
   - The overhead projector (OHP) is a device that shines light through transparent film with writing and drawings on it, projecting them onto a screen. The source says this technology was mainly used for military purposes before it entered classrooms.
   - The source says this is the final episode of the fifth season of the podcast Version History. Host David Pierce, along with Jake Kastrenakes of The Verge and Stanford University professor Antero Garcia, discusses the development of the OHP and how it spread through classrooms.
   - The panelists are introduced as discussing why the OHP disappeared and whether PowerPoint, computer screens, and smart boards are better than the OHP. These are questions the hosts raise; the source does not give their conclusions.
   - At the bottom of the source are additional reading links, including a history of the magic lantern, the history of innovation at 3M, and an NPR article on American science education after Sputnik.
205. [Running your own intermediate server for AI model calls lets you see usage records](https://jrdsgl.com/if-you-dont-run-your-own-llm-proxy-how-the-fuck-do-you-even-know-whats-going-on/) (jrdsgl.com)
   0 points by startupxo, 2026-10-19T10:31:22.356Z | 0 comments | https://startupxo.com/en/1124/ | https://startupxo.com/en/1124.md
   - The intermediate server that requests pass through when calling an AI model is called an LLM proxy. The author, Jared Siegel, says that if you run this intermediate server yourself, you can check which calls went out, when, and to which model.
   - According to the original claims, one proxy can group multiple providers and models, record which key called which model, and let you use separate virtual API keys for each task. These are the author's stated advantages, and the original gives no performance figures.
   - The author says that when a cache miss occurs during coding work (a situation where a stored result cannot be used and must be recalculated), he can find the cause by hand within about 20 seconds. He has not automated this yet, but he adds that the proxy has an API, so automation is possible.
   - The author does not recommend any specific proxy product. Instead, he advises installing one under your own control and modifying it to fit your needs, and he describes adding a lookup function and an MCP server connection to his own setup.
206. [Indian hospital Even changes its operations to patient recovery outcomes instead of test and surgery counts](https://techcrunch.com/2026/10/19/khosla-ventures-doubles-down-on-even-healthcare-to-challenge-indias-hospital-economics/) (techcrunch.com)
   0 points by startupxo, 2026-10-19T08:44:21.175Z | 0 comments | https://startupxo.com/en/1123/ | https://startupxo.com/en/1123.md
   - Even is an Indian startup founded in 2021 that aims to change a model in which hospital profits grow as tests or admissions increase. The company builds its own hospitals, pays doctors fixed salaries, and said it can pay bonuses based on patient recovery outcomes rather than the number of tests or surgeries.
   - In a Series B round led by Khosla Ventures, it raised $50 million, and according to the company, its valuation is $300 million. Cumulative investment exceeds $120 million.
   - The company said that the unplanned readmission rate at its own hospitals is below 1%, lower than the industry benchmark of 8 to 15%, and that patient length of stay is 40% shorter. These are figures Even presented, and the source contains no verification from an independent body.
   - According to filings submitted to regulators, its net loss for fiscal year 2025 was about 902 million (about $9.5 million), and revenue was about 272 million (about $2.86 million). Reports from the Indian startup outlet Entrackr said the company laid off about 350 people in August.
207. [Linux 7.3 kernel officially released with record-scale changes](https://lwn.net/Articles/1101143/) (lwn.net)
   0 points by startupxo, 2026-10-19T08:43:49.128Z | 0 comments | https://startupxo.com/en/1122/ | https://startupxo.com/en/1122.md
   - The Linux kernel is the core software that connects computer hardware and programs. According to the source, Linus Torvalds officially released version 7.3, and the merge window for version 7.4 starts the next day.
   - The source says this release set records in every metric. Before the merge window, 18,765 commits came in from 3,021 developers, and 807 of them were participating in kernel development for the first time.
   - Main changes include extensions to binfmt_misc, group scheduling work aimed at improving performance on large systems, a new filesystem failfs that makes all operations fail, the fchroot() system call, and global per-CPU data for BPF programs.
   - It also includes completing sched_ext sub-scheduler features, tracking virtual machine work sets, alternate data stream support for the ntfs3 filesystem, and SMB server improvements. For more details, see two LWN merge window summaries and KernelNewbies' guide page for 7.3.
208. [superlite, an automatic build tool for Claude Code: hand over a spec and receive tested results](https://contextfirst.blog/superlite-hand-an-agent-a-spec/) (contextfirst.blog)
   0 points by startupxo, 2026-10-19T06:57:34.566Z | 0 comments | https://startupxo.com/en/1121/ | https://startupxo.com/en/1121.md
   - This tool is an extension added to the AI coding tool Claude Code. When a user hands over a spec, it reduces approval questions, multiple AI agents split and build the code, and then testing and review proceed. It is a project created by author Iheatu Chris Wogu.
   - According to the original author's benchmark, on the same project evaluated with 177 hidden tests, the average of 3 runs of the default superpowers setup was 33.6 minutes, and the latest superlite run of 1 took 16.1 minutes. The author conducted the measurement, and both approaches reportedly passed 177 of 177.
   - Work is split by file and interface, and independent tasks run at the same time. According to the original, the latest run had up to 18 agents working simultaneously. Tests run in the background so progress does not stop.
   - The original explains that rules are enforced by hooks, a code mechanism, rather than by prompt wording. It reportedly blocks the coordinating agent from directly editing project code, and records automatic decisions so they can be reviewed later.
209. [Anthropic announces plan to contribute 5% of company shares to a public benefit fund](https://techcrunch.com/2026/10/18/anthropic-says-its-committing-5-of-its-shares-to-a-public-benefit-fund/) (techcrunch.com)
   0 points by startupxo, 2026-10-19T06:56:53.967Z | 0 comments | https://startupxo.com/en/1120/ | https://startupxo.com/en/1120.md
   - Anthropic announced it will contribute 5% of its company shares to a 'public benefit fund (public benefit fund)'. According to the original report, the fund will be set up as a fully owned subsidiary inside the company.
   - The company also said it may contribute an additional 2% each time the company's value rises by 50%. This is the company's own statement.
   - Examples given for the fund's use include large donations to traditional nonprofit organizations and direct support for specific groups. The original report does not give specific support criteria.
   - The company said it has already invested $200 million in studying how AI will affect the job market. It said it will publish more details on the fund's operating structure and decision-making process within the next few months.
210. [Cyberspace trademark filing and William Gibson's objection, as the author recounts](https://colo.to/cyberspace.html) (colo.to)
   0 points by startupxo, 2026-10-19T04:35:04.850Z | 0 comments | https://startupxo.com/en/1119/ | https://startupxo.com/en/1119.md
   - Author Erik Gullichsen says he built a virtual reality prototype at the Autodesk research lab in 1989. The setup connected a head-mounted display, a position tracker, and a data glove that reads finger movements to a 386 computer, and the source says the system ran at roughly 6~10 frames per second.
   - According to the source, Autodesk filed to register 'cyberspace' as a trademark, and William Gibson sent a letter through his lawyer objecting to it. The author recalls that he did not respond to this letter.
   - According to the author, the claim that 'Gibson tried to trademark Erik Gullichsen's name' was spread as a prank by Timothy Leary. The source explains that Paulo and Rudy Rucker later added to the story and spread it, and that it was eventually taken as fact.
   - At an event in Barcelona in 1990, the source reports that Gibson quoted the idea that a street finds its own uses for things and said, 'Autodesk is no street.'

Next page: https://startupxo.com/en/new.md?p=8
JSON: https://startupxo.com/api/v1/new.json?lang=en&page=7
Other languages: https://startupxo.com/ko/new.md https://startupxo.com/ja/new.md https://startupxo.com/es/new.md https://startupxo.com/ar/new.md
