Byte Sized Security

Episode Summary:In this short episode, Marc covers 404 Media's reporting on Amanda Smith, a Pennsylvania woman arrested on an Arkansas warrant after someone opened a Cash App account with her Social Security number and date of birth and used it to collect $1,000 from a Facebook Marketplace scam. He walks through how investigators tied the account to her, the account details that don't match her, and why Cash App wouldn't share account information with her. The case is ongoing and Smith is presumed innocent. Marc closes with one record you can start keeping this week.Key Topics Covered:The arrest: a warrant from Arkansas, a computer fraud charge, and three nights in jailHow police got to her: a hacked Facebook account, a fake Marketplace sale, and a search warrant to Block, which owns Cash AppThe data that didn't match: a Philadelphia address, an iPhone running a 2022 operating system, and a Verizon connection when she's on AT&TWhere the case stands: Cash App wouldn't share account details with her, she turned down a plea offer, and a hearing is set for NovemberMain Takeaways:Identity theft isn't only a money problem. Stolen identity data can attach your name to someone else's crime.In this case, a Social Security number and date of birth were enough to put Smith's name on a payment app account.The details that didn't line up were the address, the device, and the phone carrier.Write down which accounts are actually yours: your banks and payment apps, your phone number, your carrier, and your devices.Timestamps:[0:00] The hook[0:17] The arrest[0:35] How police got to her[1:06] The data that didn't match[1:27] Where the case stands[1:50] What you can do this weekTools & Resources Mentioned:404 Media, Joseph Cox, "Someone Opened a CashApp In Her Name. She Faces Years in Prison" (Oct 21, 2026): https://www.404media.co/amanda-smith-opened-a-cashapp-in-her-name-she-faces-years-in-prison-cashapp-fraud/FTC IdentityTheft.gov, report identity theft and get a recovery plan: https://www.identitytheft.gov/FTC, Credit Freezes and Fraud Alerts: https://consumer.ftc.gov/articles/what-know-about-credit-freezes-fraud-alertsFTC, What To Know About Identity Theft: https://consumer.ftc.gov/articles/what-know-about-identity-theftEquifax credit freeze: https://www.equifax.com/personal/credit-report-services/credit-freeze/Experian credit freeze: https://www.experian.com/freeze/center.htmlTransUnion credit freeze: https://www.transunion.com/credit-freezeSocial Security Administration, Fraud Prevention and Reporting: https://www.ssa.gov/fraud/Free credit reports: https://www.annualcreditreport.com/This episode's audio was produced with an AI voice clone of Marc, from a script Marc reviewed and approved.Not legal advice. If you're facing criminal charges, talk to a licensed attorney.---I do hope you enjoyed this episode of the podcast. Here are some helpful resources including any sites that were mentioned in this episode.--Listen and subscribe: https://bytesizedsecurity.show/listen--Support ByteSizedSecurity (501(c)(3)): Donate via GiveLively--Questions or tips: email [email protected] (subject: Byte-Sized Security). TikTok: @bytesizedsecurity

Episode Summary:ChatGPT now shows ads on its Free and Go plans. In this episode Marc walks through what 404 Media found when it ran its own test ads: advertisers can upload lists of emails, phone numbers, or Android advertising IDs, and your current chat shapes which ad appears. Turning off ads personalization changes how ads are picked, but it does not stop them. Marc covers what each setting does and five steps you can take today.Key Topics Covered:404 Media's test ads: how Joseph Cox used OpenAI context hints to steer ads toward privacy and AI chatsCustom audiences: email, phone, hashed identifiers, and Google Advertising IDs, plus include, exclude, and bid-adjustment bucketsThe two switches: Personalize ads vs Ads-Free mode, and why one does not equal the otherWhat OpenAI commits to: advertisers do not get your chats; Temporary Chats show no ads; health and politics limitsFive actions: check your plan, Data Controls, delete ads data, Temporary Chat, and shrink the lists you are onMain Takeaways:Free and Go show ads. Plus, Pro, Business, Enterprise, and Edu do not.Context hints are guidance, not keyword triggers, but your prompt still decides whether an ad fits.Advertisers can match uploaded emails, phone numbers, or Android advertising IDs to ChatGPT users (inclusion and bid bumps need at least 25,000 matched users).Turning off ads personalization still leaves context-based ads on. Ads-Free (Free plan only) or a paid plan is what removes ads.Your personalization setting does not take you off an advertiser's uploaded list.For sensitive topics, use Temporary Chat. It shows no ads and does not use or update memory.Timestamps:[0:00] Cold open[0:50] Intro[1:13] What 404 Media did[2:22] Your email and phone number[4:17] The switch which doesn't stop ads[5:42] What OpenAI promises[6:27] Five things to do today[7:42] OutroTools & Resources Mentioned:404 Media (Joseph Cox, Oct 15, 2026): How ChatGPT Targets Ads To You Based On Your Prompts, Phone Number, and EmailOpenAI Help: Ads in ChatGPTOpenAI Help: Set up Custom Audiences for your CampaignOpenAI Help: Write Context Hints for ChatGPT AdsOpenAI: Our approach to advertising and expanding access to ChatGPTGoogle: Advertising IDByteSizedSecurity blog: ChatGPT ads targeting (prompts, email, phone)Not legal advice. Settings and plan details reflect OpenAI's help pages as of October 15, 2026 and change over time.---I do hope you enjoyed this episode of the podcast. Here are some helpful resources including any sites that were mentioned in this episode.--Listen and subscribe: https://bytesizedsecurity.show/listen--Support ByteSizedSecurity (501(c)(3)): Donate via GiveLively--Questions or tips: email [email protected] (subject: Byte-Sized Security). TikTok: @bytesizedsecurity

A 20-second poem about multi-factor authentication (MFA), the second step that proves it's really you when you log in.Main TakeawayA password alone can be copied, lost, or sold. Add a second factor, ideally a hardware security key or passkey, which a fake login page can't capture.Timestamps0:00 The poemTools & Resources MentionedCISA: More than a Password

Episode Summary:Someone told Marc that AI panic is nothing new — just the printing press or nuclear weapons all over again. He disagreed, and it turns out there was a report to back it up. In this episode he breaks down why AI collapses the cost of dangerous capability in a way the printing press, the internet, and even nuclear weapons never did, what Anthropic's brand-new September 2026 threat intelligence report documents, where his own "$200 expert" framing overstated the case, and the four guardrails that would close the gap.Key Topics Covered:The argument that started this episode — a debate about whether AI panic is history repeating, and the report Marc found three days later making his case for himWhy the printing press comparison breaks down — institutions had a century (and decades, for the internet) to catch up; AI's capability curve moves in monthsWhat's actually different about nuclear weapons — nuclear risk lives behind physical choke points: materials, facilities, expertise. AI risk lives in a skill, and skills can't be fenced offAnthropic's report: the receipts — three disrupted operations, walked through case by case, that turn the argument from speculative to documentedDoes AI make anyone an expert? Not exactly — Marc's own "$200 subscription = expert" line, and the more defensible version of the claimAttackers, defenders, and who adapts faster — the same models cutting attacker costs are cutting defender costs too, and why that race mattersWhat real AI guardrails would look like — four concrete guardrails: pre/post-release capability testing, enforceable standards, international coordination, and risk-scaled accessMain Takeaways:AI doesn't need generations to reach scale like the printing press or the internet did — model capability jumps happen every few months, not every few decadesNuclear risk is contained by physical choke points (fissile material, facilities, expertise); AI risk lives in a skill, and skills don't have a border to fenceAnthropic's September 2026 report documents real, disrupted operations — including a breach that went from one stolen developer token to full cloud admin control in roughly three hours"$200 subscription = expert" overstates it: AI doesn't manufacture expertise, it lowers the skill required to attempt tasks whose consequences the operator isn't trained to handleDefenders get the same acceleration attackers do — the organizations lagging on AI-assisted defense are the ones absorbing the most riskClosing the gap takes four things: pre/post-release capability testing, enforceable (not voluntary) standards, international coordination, and access that scales with risk instead of priceTimestamps:[0:00] The argument behind this episode[1:03] Why the printing press comparison breaks down[1:53] What's different about nuclear weapons[2:55] Anthropic's report: the receipts[4:46] Does AI make anyone an expert? Not exactly[5:45] Attackers, defenders, and who adapts faster[6:16] What real AI guardrails would look likeTools & Resources Mentioned:Anthropic: Detecting and Countering Misuse of AI (September 2026)NIST AI Risk Management FrameworkEU AI Act (European Commission)Full written guide: Why AI Risks Are DifferentAI is fueling the cybersecurity career boomWhy an AI agent shouldn't inherit your permissionsNot legal advice. Figures reflect Anthropic's report as published on September 10, 2026.---I do hope you enjoyed this episode of the podcast. Here are some helpful resources including any sites that were mentioned in this episode.--Find subscriber links on my site, add to your podcast player, or listen on the web players on my site:Listen to Byte Sized Security --Support this Podcast with a Tip:Support Byte Sized Security --If you have questions for the show, feedback or topics you want covered. Please send a short email to [email protected] with the Subject line of "Byte-Sized Security" so I know it's about the podcast.Connect with me on TikTok: https://www.tiktok.com/@bytesizedsecurity

Roughly 98.5% of all known CVEs have never been exploited. In this episode I break down a conversation between Jeremiah Grossman and Robert Hansen of Root Evidence, and host Raphael Mudge, on the Down the Rabbit Hole podcast, and what it means for how you prioritize a patch queue. I cover CVSS score versus exploitation evidence, how to use CISA's free KEV catalog, why the vulnerability management industry has no incentive to tell you the truth, and what separates a junior-sounding answer from a senior one in a security interview.In this episode:(00:00) The scan report that isn't as urgent as it looks(01:03) The 98.5% number and the mechanic analogy(02:01) Why the industry defaulted to patch everything(03:00) The 36-hour outage from a perfect-10 patch(03:48) CVSS score vs. exploitation evidence vs. insurance-claims data(05:08) What it sounds like when someone understands this in an interview(06:15) Why the industry has no brakes, and the AI-hype myth(07:49) Your homeworkLinks:Down the Rabbit Hole, episode 722, "Vulnerability Math Ain't Mathing"CISA's Known Exploited Vulnerabilities (KEV) catalogFIRST.org, the CVSS specificationFull written breakdownOur cybersecurity career guideBreaking into cybersecurity with no experienceThird-party risk and the AI bug-report floodNot financial or legal advice. Figures cited reflect Root Evidence's analysis as discussed on the source podcast episode.I do hope you enjoyed this episode of the podcast. Here are some helpful resources including any sites that were mentioned in this episode.--Find subscriber links on my site, add to your podcast player, or listen on the web players on my site:Listen to Byte Sized Security --Support this Podcast with a Tip:Support Byte Sized Security --If you have questions for the show, feedback or topics you want covered. Please send a short email to [email protected] with the Subject line of "Byte-Sized Security" so I know it's about the podcast.Connect with me on TikTok: https://www.tiktok.com/@bytesizedsecurity