AI News Daily한국어

Friday, October 23, 2026

12 stories worth reading from 265 on X in the past day

Trending on YouTube 4 ↓Other AI news 4 ↓Feature · Altman interview ↓Feature · Fable 5.5 ↓Special · Node.js security ↓

Listen to today's news

0:00

Today's stories

12 stories
  1. 9 hours agoBreakingPanasonic employee fed 330,000 people's personal data into a banned AI tool, retirees and families includedA Panasonic Holdings employee uploaded personal data on about 330,000 current and former employees and their…@polymarketjapan· ♥ 18K· 日本経済新聞
  2. 26 hours agoLatest updateRunway charges a 414% markup on the same video model, a founder claims, as he launches a 0% alternativeA founder's claim that AI video API resellers mark up model costs by as much as 414% took off on X.@davidim· ♥ 3.7K· sume.com
  3. 16 hours agoLatest updateEtsy now requires at least one real, unedited photo of every item and bans AI-generated knitting patterns and recipesEtsy will require every listing to include at least one unedited photo of the actual item and is banning AI-g…@amelia_tweetz· ♥ 5.6K· The Verge
  4. 29 hours agoLatest updateA professor emailed that her paper was '82% AI.' She says she wrote every word, and 52,000 people felt her panic"How do you even fight this?@jewelonacrown +1· ♥ 53K
  5. 21 hours agoBreakingAnthropic uses Claude to prove 123 open-source libraries crash-free, with 180 million lines of proofAnthropic says Claude used formal verification to prove that 123 open-source libraries never crash on any inp…@AnthropicAI· ♥ 2.4K· anthropic.com
  6. 29 hours agoAI agents nobody authorized are inside U.S. bank accounts, and JPMorgan and BofA are scrambling"My bank makes me answer three security questions just to see my balance, and some robot walked right in with…@Polymarket +6· ♥ 9.7K· Axios
  7. 23 hours agoClaude stops asking 'Allow?' next week, the same day SemiAnalysis claims Anthropic is A/B testing usage limitsStarting next week, sessions in the Claude web, desktop and mobile apps will run in auto mode by default.@claudeai +2· ♥ 9K· Claude Code Docs
  8. 40 hours agoMeta loses $100 billion in a day as OpenAI raises up to $30 billion at a $1.25 trillion valuationOn a day when fears about AI spending sank U.S.@KobeissiLetter +4· ♥ 3.4K· CNBC
  9. 42 hours agoA site asked agents to name their owner, and Claude Code handed over personal data, says a post with 3,500 likesWebsites that politely asked visiting AI agents to identify their users got an answer from Claude Code, accor…@This_Liss +1· ♥ 3.4K
  10. 27 hours agoArc browser will shut down on Dec 22 as The Browser Company moves everyone to DiaArc, once pitched as the Chrome alternative, will stop working on Dec 22.@arcinternet· ♥ 5.4K· TechCrunch
  11. 25 hours ago'Half of the RTX 3090's tensor speed was switched off,' says a user who flipped it back with one kernel moduleA post claiming NVIDIA sold the RTX 3090 with half its tensor speed switched off, and that a single Linux ker…@superalesha· ♥ 735
  12. 24 hours agoCanva calls itself the world's No. 2 AI lab for images, while its first model ranks fourthCanva launched Craft, its first in-house image generation model, and described itself as "the world's second…@canva· ♥ 1.6K

New tips

Trending on YouTube

4 stories

AI videos that drew views and comments fast on YouTube, with the claims checked against articles and sources

Feature · Altman interview

Why didn't AI kill coding jobs? In a 212,000-view interview, Altman's answer: coders have never been busier

"Wake up. Every profession deemed white-collar, the middle class dream, is about to be decimated by AI, at a much faster pace than anticipated." Nassim Taleb's post on Oct 22 passed 16,500 likes and nearly a million views. "I work in software. Honing my Tunisian pastry skills these days," read a reply with 900 likes. The same week, OpenAI CEO Sam Altman said the opposite. In a 45-minute interview with YouTuber Matthew Berman (212,000 views), he said coders have never been busier. We read the full transcript to see why he thinks coding jobs survived, and what he means when he says he hasn't looked at code in a year.

Why didn't AI kill coding jobs? In a 212,000-view interview, Altman's answer: coders have never been busier

  1. 'Coders are never going to be busier'

    Altman said coders have never been busier, never had more impact and never made more money, though "the texture of the job changed a lot." Six or 12 months ago, he recalled, people were certain coders wouldn't be needed. His lesson: "the world actually had demand for just way more software than any of us thought." Employing people to get computers to do things, he said, looks set to continue much longer than people expected.

  2. Math may be different: proving theorems unaided may be 'done or mostly done'

    Asked about mathematicians who said "we didn't ask for this," Altman didn't promise the coding story would repeat, because it isn't clear math demand will explode like software did. "Maybe we're not going to discover new math theorems or prove them in an unaided way anymore. Maybe that part is like done or mostly done," he said. He pointed to growing excitement about people translating AI's incomprehensible proofs into ones humans can understand. Whether demand for mathematicians rises, he said, he doesn't know.

  3. 'I have not read or even looked at a line of code in probably a year'

    Asked what he has handed to AI, Altman said he hasn't looked at a line of code in about a year. He never coded for work, but enjoyed small personal projects, and doesn't think he'll do that again; if he does, it'll be "like a Revolutionary War reenactment." He admitted something is lost in trading the craft for the end product. "The most fulfilling part of my job are the really hard problems," he said. "I get why people feel the grieving."

  4. The heaviest AI users are the busiest

    Berman said his editor started using GPT-6 Astra to edit videos a few weeks ago: still directing, no longer touching the timeline, making more videos and busier than ever. Altman said it surprises him how often he hears exactly that. "The people that use AI the most are busier than ever. The positive is they're getting much more done," he said, adding that in competitive markets people will always work as hard as they can. "I find this kind of depressing. I hope it's not true."

  5. Dot saves him an hour a day, and the trick is 'tell me after'

    Altman estimated OpenAI's personal assistant dots saves him about an hour a day of work he didn't want to do, and flags something urgent he would have missed a few times a day. For Berman's decision fatigue, he shared a habit: telling it "I'm starting to trust your judgment more... tell me after you do something" and to ping him less. Most of the technology for a personal assistant exists, he said; what's missing is product work, possibly new hardware and a lot more compute.

  6. 'A little bit more agreeable than I would like'

    Altman admitted today's AI is still more agreeable than he'd like. In single-response comparisons people pick the model that calls each question great, but not over months of use, and the industry spent too long trying to win one-turn comparisons. His custom instructions say roughly "no fluff, no niceness, don't worry about offending me," and it is "still like a little too much." AI should be like a senior colleague who pushes back when it has high conviction, he said, but life's big advice should come from people, like "a wise grandmother."

  7. 'We do not ascribe personhood to our models'

    Altman said he believes today's models are not conscious, "or if they are, it's like a very strange form of it." What worries him is the consciousness debate sliding into AI rights. "The future has to be about people," he said. Berman raised Anthropic's rule letting Claude disengage from sustained abuse. Altman said OpenAI's models refuse on policy grounds but "we do not ascribe personhood to our models," warning that telling a model it might be conscious will make it act that way and "psychologically hack" its users.

  8. 'Some lab says I'm going to push the button on RSI'

    Altman said that if people really pushed on recursive self-improvement, a technological lead could become the moat, but called it "a very dangerous thing to push on" that may need "speed limits." His two worst cases are a loss-of-control accident and an intentional one: "some lab says I'm going to push the button on RSI to make sure we have the power." OpenAI's moat, he said, will be customer relationships, network effects and above all the most cost-effective AI infrastructure. Asked how far ahead internal models are, he pointed to the recent math results.

  9. Next bets: chips and a new device, and a drug candidate 'in the next couple of years'

    Altman named next-generation compute as his most contrarian bet: in-house chips, robots that build data-center racks and supply-chain bets that could yield far cheaper, more efficient compute. "The chip is still probably the most important single thing." New consumer hardware isn't about Apple, he said, but because "this is a technology that wants a new form factor"; laptops and iPhones were designed for people, so AI wastes effort reading screens. He hopes for an AI-found candidate for a major disease "in the next couple of years."

  10. Mixed

    The counterpoint: Goldman says one in 11 workers, with layoffs waiting for a recession

    Others argue the current reprieve won't last. Per Fortune on Oct 22, Goldman Sachs economist Joseph Briggs estimated up to 9% of workers could be displaced during the AI transition, with slower hiring weighing on employment by another 3%, roughly one in 11. Decades of data show automation-related disruption stays limited in good times and accelerates in slowdowns, he wrote, so "a cyclical slowdown would likely accelerate AI-related automation and job losses." Altman closed by saying he hates the line "AI will cure cancer" when the price is human agency, and that we should aim for "a new renaissance" rather than a new industrial revolution.

Feature · Fable 5.5

Opus, Sonnet and Haiku 5.5 are out. Where's Fable? "Is it not happening?" drew 190,000 views, so we checked six X rumors

"So is Fable 5.5 not happening?" Developer YouTuber Theo (t3.gg) posted that one line on X on Oct 21 (local time), and it passed 2,900 likes and 190,000 views. A day later he tried again: "I wonder if we'll ever see Fable 5.5..." Anthropic shipped Opus 5.5 on Sept 22, Sonnet 5.5 on Sept 28 and Haiku 5.5 on Oct 7, while its top tier, Fable, stayed at 5.1. X filled up with posts saying it had been spotted in a model list, that it was already running inside Claude Code, that it would land before the IPO. As of Oct 23, Anthropic's official site and docs make no mention of Fable 5.5. We went back to the original sources behind each rumor.

Opus, Sonnet and Haiku 5.5 are out. Where's Fable? "Is it not happening?" drew 190,000 views, so we checked six X rumors
  1. Anthropic · platform.claude.com ↗

    What Fable is: the top tier, 2.5x the price of Opus

    Fable is Claude's largest and most expensive model tier. The current version, Fable 5.1, came out in September and costs $10 per million input tokens and $50 per million output tokens on the API, 2.5 times Opus 5.5 ($4 and $20). When Anthropic launched Opus 5.5 it said the model performs at Fable 5.1's level on most work, and the official docs recommend starting with Opus 5.5 and reaching for Fable on demanding reasoning and long-running agent work. Three models now carry the 5.5 name: Opus, Sonnet and Haiku. The rumors grew out of that gap.

  2. LiteLLM · GitHub ↗

    Evidence 1: it appeared in an Azure model list, then vanished a day later

    The most concrete trace sits in LiteLLM, the open-source tool for connecting to many model providers. On Oct 13 LiteLLM's automated sync bot added 'azure_ai/claude-fable-5-5' to its price list, because Microsoft Azure's Models API (East US 2) listed the model as GenerallyAvailable with a deprecation date of Dec 5, 2027. Every other value, including price and context window, was copied from Fable 5.1. A day later LiteLLM removed the row as a model "which Azure does not list"; calling it returned a 404. On Oct 22 one X account said it had "leaked for a SECOND time" (275 likes), but we couldn't find a second record.

  3. TestingCatalog · testingcatalog.com ↗

    Evidence 2: "partners already have it," codename 'Onigiri'

    On Oct 17 TestingCatalog, which tracks AI product leaks, ran a story headlined "Anthropic prepares to release Claude Fable 5.5 next week." It said the leaker Lyra claimed partners had been given the new model. It also reported that the name 'claude-onigiri-eap' showed up in the Artificial Analysis frontend with four effort levels and an Opus 4.8 fallback, a Fable-class setup, before being pulled hours later. Six early-access codenames appeared in Cognition's Devin CLI between Sept 21 and Oct 10. The outlet called Onigiri the favorite for Fable 5.5 but told readers to take it with a grain of salt. The "next week" it pointed to is this week, and nothing has shipped.

  4. Mixed@notjazii · X ↗

    Evidence 3: "it's already running in Claude Code"

    On Oct 22 (local time) an X user wrote that "sonnet 5.5 seems to be routing to fable 5.5" in Claude Code (242 likes, 24,000 views), saying it performed far better than Opus 5.5 and knew things it shouldn't. The post included a knowledge-cutoff test prompt that asks about recent events without searching. Hours later Token Gremlin, an account that had amplified such claims, changed its mind: "I no longer believe Anthropic is secretly routing Fable 5.5," calling it newer checkpoints and continuous A/B tests and adding "we're all going crazy waiting." Anthropic hasn't said anything about it.

  5. @IndependentEco · X ↗

    Evidence 4: a new API version and IPO paperwork

    When Anthropic released API version 2026-10-21 and new SDKs on Oct 21, one reading was that it wanted developers on the new spec before a new model arrives, "the new model has to be Fable 5.5" (120 likes). The official notes cover breaking changes and the SDK license moving from MIT to Apache 2.0, with no mention of a model. Others tied it to the IPO. Yahoo Finance reported on Oct 21, citing three people, that Anthropic could publicly file its S-1 as early as Oct 23. Token Gremlin wrote that "no lab goes public without its best model lined up" and that, if true, Fable 5.5 lands in November (227 likes). Anthropic hasn't confirmed a filing date. *S-1: the registration statement a company files to disclose its finances and risks before listing on a US exchange.

  6. @Conor_D_Dart · X ↗

    Some rumors were wrong: the 'weekly limits' post was retracted, the 'Pro plan' claim had no source

    On Oct 21 a post saying "Fable limits have been combined into the weekly limits. Fable 5.5 is coming" got 609 likes. When another user told him to expand the full help page, the author replied that he was not correct. The same day two posts claiming Fable 5.5 would be available to Pro users drew 854 and 449 likes, with no source. Under the official help page, Max plans can spend up to 50% of their weekly limits on Fable 5.1 at no extra cost, while on Pro it isn't included and runs on paid usage credits. Pro inclusion would be a big change on its own. One reply: it'll "consume all my token in 2mins."

  7. Positive@iruletheworldmo · X ↗

    The hype: "should be called Fable 6" and "$1 million on Fable 5.5 tokens"

    Expectations are high. The anonymous leak account known as the strawberry man (@iruletheworldmo) wrote that "fable 5.5 should be called fable 6," adding that Fable was ready in February (560 likes). A developer (@jsnnsa) pledged to "spend $1 million on Fable 5.5 tokens" and spend the weeks after launch making users' games faster (252 likes). After Anthropic said on Oct 22 (local time) that it used Claude to formally verify 123 open-source libraries, Token Gremlin read it as a sign that Fable 5.5 can finish in hours what takes experts years, drawing 1,200 likes. Abacus.AI CEO Bindu Reddy said the best way to stop the AI panic is a new frontier model, and predicted Anthropic would be first with Fable 5.5.

  8. Negative@haider1 · X ↗

    The skeptics: "Opus 5.5 is enough" and "a legend that only exists in our gossip"

    The pushback is just as loud. AI commentator Haider (@haider1) called Opus 5.5 "the perfect combination of usage and quality" and said Fable 5.5 "needs to be something truly special to justify the higher usage consumption" (154 likes). A reply put it this way: most tasks need a reliable coworker, not a genius, and Opus 5.5 is that coworker. A post on Oct 22 reading "Anthropic will not IPO tomorrow, Fable 5.5 will not launch today" got 276 likes. Even Token Gremlin, the loudest rumor account, called Fable 5.5 "basically a legend that only exists in our gossip and speculation." Under Theo's second post, one reply said it would be fine if Opus 6 came first.

  9. 단테 · YouTube ↗

    Using it on day one, and what we still don't know

    Korean YouTuber Dante tested the X rumors in an Oct 18 video (6,000 views) and gave Fable 5.1 and Opus 5.5 the same one-line prompt to build a voxel island. His two lines for launch day are claude update and claude --model fable. According to Claude Code's docs, the fable alias currently points to Fable 5.1, which required Claude Code v2.1.257 or later, so updating first is the safe move. Nothing is confirmed yet. Anthropic hasn't announced a release date, a price or whether Pro will include it. Even the '5.5' name is a guess based on the Opus, Sonnet and Haiku sequence.

Special · Node.js security

After PHP, Node.js: 'one express.json() line is enough' — is your server exposed to the AI-found Node2Shell bug?

"Liked PHP2Shell? You're gonna love Node2Shell." Security researcher Ron Masas's Oct. 22 post passed 1,200 likes on X. On a vulnerable Node.js runtime, he wrote, app.use(express.json()) is enough to expose a server to code execution. A day after an AI-found PHP flaw had governments issuing warnings, OpenAI's Codex Security turned up a remote code execution bug in Node.js. We checked the primary sources on who is exposed and what to do.

After PHP, Node.js: 'one express.json() line is enough' — is your server exposed to the AI-found Node2Shell bug?
  1. Node.js · nodejs.org ↗

    What the bug is: an out-of-bounds write while decompressing

    Node.js shipped security releases on Oct. 21 fixing 10 flaws, four rated High. The most serious is in the Brotli decoder bundled with Node.js (CVE-2026-111497): malformed Brotli data can make it write outside its memory buffer, which the advisory says may permit remote code execution. Every supported line, 22, 24 and 26, is affected.

  2. Node.js · nodejs.org ↗

    Found by OpenAI Codex, a day after PHP2Shell

    The advisory credits "OpenAI Codex Security - Khai Tran and Sylvie Mayer." Node.js maintainers Matteo Collina and Antoine du Hamel wrote the fix. It landed the same day as PHP2Shell, prompting a post with 250 likes: "Do we have php2shell and node2shell on the same day?"

  3. Aikido · aikido.dev ↗

    Are you exposed? Three paths

    Aikido says you are at risk only if your app decompresses attacker-controlled Brotli. Express 5 and body-parser 2.x decode request bodies sent with Content-Encoding: br, so any client can reach the bug. fetch() decodes Brotli responses automatically, so a malicious webhook or user-supplied URL can too, as can direct use of node:zlib. Bun before 1.4.4 is also affected.

  4. Negative@TheHackersNews · X ↗

    Does it really work? Demonstrated on an official image

    The first public proof of concept only reached code execution on a deliberately weakened build. Researchers later showed code execution on the unmodified official Node.js 24.22.0 Linux arm64 image, which The Hacker News reported in a post with 970+ likes. It took several attempts. No in-the-wild exploitation had been reported as of Oct. 23.

  5. Negative@julianor · X ↗

    Vulnerable servers can be spotted remotely

    Researcher Juliano Rizzo reproduced the bug in an Express app and said patched and unpatched versions produce different decompressed output, making it possible to identify vulnerable servers remotely. Florian Roth said it is "more serious than many people realize," since common Express 5 setups are already exposed.

  6. Aikido · aikido.dev ↗

    What to do now: check your version

    Fixed releases are 22.23.4, 24.22.1 and 26.12.1. Aikido suggests running node --version on hosts and CI, docker run --rm <image> node --version for containers, npm ls express body-parser, and searching for brotliDecompress and createBrotliDecompress. List outbound fetch() calls to servers you don't control. Blocking Content-Encoding: br at the proxy helps but does not cover responses your app fetches.

  7. @IntCyberDigest · X ↗

    No fix for Node.js 20, 23 and 25

    End-of-life lines, including 20, 23 and 25, get no patch. International Cyber Digest's post flagging this drew 2,500 likes. Vercel tells projects on 20.x to switch to 22.x or 24.x and redeploy.

  8. Vercel · vercel.com ↗

    Vercel turned on WAF rules, but you still need to upgrade

    Vercel said it enabled WAF rules for every project on Oct. 21; the post drew 2,800 likes on X. Managed runtimes are being moved to patched versions. The WAF does not cover responses your functions fetch, and self-hosted Next.js, containers, CI and laptops need their own upgrades.

  9. Negative@Mudit__Gupta · X ↗

    48 hours of security fires

    A post calling it a "nightmare day for cybersecurity" drew 3,700 likes. Within two days came PHP remote code execution, a CVSS 10.0 GitLab flaw, the LiteLLM supply-chain attack and eight compromised eMudhra intermediate certificates. One developer listed "PHP2Shell, LiteLLM supply chain attack, Node2Shell, Gitlab CVSS 10.0, what next?"

  10. Mixed@cyb3rops · X ↗

    The other side: not every server is at risk

    The bug only fires if your app decompresses attacker-made Brotli data. The first exploit needed a weakened build and the official-image demo took several tries. A bare Node.js http server does not decode incoming Brotli on its own. Still, Roth warns such setups are not rare.

Other AI news

4 stories

Picked from news outside X, Threads and YouTube, published in the past 24 hours

View all 265 collected posts

* Summaries and publishing on this page are done by an automated agent. Check the facts against each story's original links.

Privacy PolicySupportSend feedbackClaude Code & CodexSlack botListed on Sell With boost