{"person_view":{"person":{"id":18039211,"name":"danzabia","display_name":"danzania","banned":false,"published":"2025-06-01T01:52:02.476822Z","actor_id":"https://infosec.pub/u/danzabia","local":false,"deleted":false,"bot_account":false,"instance_id":86},"counts":{"person_id":18039211,"post_count":1,"comment_count":14},"is_admin":false},"site":{"id":73,"name":"Infosec.Pub","sidebar":"To support infosec.pub, please consider donating through one of the following services: \n\nPaypal: jerry@infosec.exchange\n\nKo-Fi: [https://ko-fi.com/infosecexchange](https://ko-fi.com/infosecexchange)\n\nPatreon: [https://www.patreon.com/infosecexchange](https://www.patreon.com/infosecexchange)","published":"2023-06-29T05:32:27.760426Z","updated":"2025-05-05T01:20:22.152468Z","actor_id":"https://infosec.pub/","last_refreshed_at":"2025-06-10T04:06:56.281712Z","inbox_url":"https://infosec.pub/inbox","public_key":"-----BEGIN PUBLIC KEY-----\nMIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8AMIIBCgKCAQEAnXWRO7espQZpFNw/uZKc\n1Q7SDuQKoooxQvomr6HH1Xpct7vNT5SHOf1ETsvlj1pMHz/tB+VutGR/mPiSprJn\n288Og43DRUjYG3YJUX94Msk8lyyJ8RFLbf/qjI+VOPYMU5+Kfr9BVGeTMyLH3aML\nPSOAVESheRmzqq13tk/XYP6d5EISUVpGJ8MIpnyyDOQSIv2J85m40UnNntAqWjO4\nuTuAv0XqTKrhw8ZHwWskHs0+3OLYWoAIpRNgjVBxwcbsHPMcPARULBjo7moiWRgB\nabMbFrgzq2borSBiC7+ilKEhTMIaTZpJDB+OauJ7AvAGvgP8Pk1A2fU0sBa6VKG8\n0wIDAQAB\n-----END PUBLIC KEY-----\n","instance_id":86},"comments":[{"comment":{"id":17282512,"creator_id":18039211,"post_id":23724289,"content":"I've been running Gemma3 4b locally on ollama and it's useful. I'm thinking about applications where a multimodal model could receive video or sensor feeds (like a security can, say).","removed":false,"published":"2025-06-10T07:12:24.471626Z","deleted":false,"ap_id":"https://infosec.pub/comment/16392124","local":false,"path":"0.17248272.17249189.17262766.17264006.17282512","distinguished":false,"language_id":0},"creator":{"id":18039211,"name":"danzabia","display_name":"danzania","banned":false,"published":"2025-06-01T01:52:02.476822Z","actor_id":"https://infosec.pub/u/danzabia","local":false,"deleted":false,"bot_account":false,"instance_id":86},"post":{"id":23724289,"name":"ChatGPT used to disable SecureBoot in locked-down device – modded BIOS reflash facilitated fresh Windows and Linux installs","url":"https://www.tomshardware.com/tech-industry/artificial-intelligence/chatgpt-used-to-disable-secureboot-in-locked-down-device-modded-bios-reflash-facilitated-fresh-windows-and-linux-installs","creator_id":765,"community_id":4523,"removed":false,"locked":false,"published":"2025-06-08T19:21:56.337371Z","deleted":false,"nsfw":false,"embed_title":"ChatGPT used to disable SecureBoot in locked-down device – modded BIOS reflash facilitated fresh Windows and Linux installs","embed_description":"A new way to fix locked firmware issues with used hardware, perhaps.","thumbnail_url":"https://cdn.mos.cms.futurecdn.net/Lad7ZdqRPXC8m33tCmvQkH.jpg","ap_id":"https://lemmy.world/post/31042464","local":false,"language_id":0,"featured_community":false,"featured_local":false,"url_content_type":"text/html; charset=utf-8"},"community":{"id":4523,"name":"cybersecurity","title":"cybersecurity","description":"An umbrella community for all things cybersecurity / infosec. News, research, questions, are all welcome! \n\n## Community Rules\n- Be kind\n- Limit promotional activities\n- Non-cybersecurity posts should be redirected to other communities within infosec.pub.\n\nEnjoy! ","removed":false,"published":"2023-06-12T13:26:10.224238Z","updated":"2024-05-24T15:48:04.914907Z","deleted":false,"nsfw":false,"actor_id":"https://infosec.pub/c/cybersecurity","local":false,"icon":"https://infosec.pub/pictrs/image/ffa014c8-749c-4482-8599-b4be5064f8b1.png","banner":"https://infosec.pub/pictrs/image/4c185ff5-b779-480f-b88e-1d922c096ba9.png","hidden":false,"posting_restricted_to_mods":false,"instance_id":86,"visibility":"Public"},"counts":{"comment_id":17282512,"score":2,"upvotes":2,"downvotes":0,"published":"2025-06-10T07:12:24.471626Z","child_count":0},"creator_banned_from_community":false,"banned_from_community":false,"creator_is_moderator":false,"creator_is_admin":false,"subscribed":"NotSubscribed","saved":false,"creator_blocked":false},{"comment":{"id":17282428,"creator_id":18039211,"post_id":23728774,"content":"Students are now prompting the AI to make it sound like a student wrote it, or putting it through an AI detector and changing the parts that are detected as being written by AI (adding typos or weird grammar, say). Even kids who write their own papers have to do the latter sometimes.","removed":false,"published":"2025-06-10T07:04:24.056493Z","deleted":false,"ap_id":"https://infosec.pub/comment/16392061","local":false,"path":"0.17251529.17252948.17258125.17282428","distinguished":false,"language_id":37},"creator":{"id":18039211,"name":"danzabia","display_name":"danzania","banned":false,"published":"2025-06-01T01:52:02.476822Z","actor_id":"https://infosec.pub/u/danzabia","local":false,"deleted":false,"bot_account":false,"instance_id":86},"post":{"id":23728774,"name":"Welcome to Campus. Here’s Your ChatGPT.","url":"https://www.nytimes.com/2025/06/07/technology/chatgpt-openai-colleges.html","body":"Link to the article without the paywall\n\nhttps://archive.ph/h63Dp","creator_id":10585107,"community_id":15,"removed":false,"locked":false,"published":"2025-06-08T21:40:33.394164Z","deleted":false,"nsfw":false,"embed_title":"Welcome to Campus. Here’s Your ChatGPT.","embed_description":"OpenAI, the firm that helped spark chatbot cheating, wants to embed A.I. in every facet of college. First up: 460,000 students at Cal State.","thumbnail_url":"https://static01.nyt.com/images/2025/06/01/multimedia/OPENAI-EDUCATION-1-05-tqlb/OPENAI-EDUCATION-1-05-tqlb-facebookJumbo.jpg","ap_id":"https://lemmy.zip/post/40766189","local":false,"language_id":37,"featured_community":false,"featured_local":false,"url_content_type":"text/html; charset=utf-8"},"community":{"id":15,"name":"technology","title":"Technology","description":"This is a [most excellent](https://media.giphy.com/media/v1.Y2lkPTc5MGI3NjExN3N0NmhuODNib3d3Nzg0OHU2bTFqMXAzNW42Y2JsOTVmenNsNG8ycSZlcD12MV9pbnRlcm5hbF9naWZfYnlfaWQmY3Q9Zw/l46CDHTqbmnGZyxKo/giphy.gif) place for technology news and articles.\n \n---\n \n## Our Rules\n \n---\n \n1. Follow the [lemmy.world rules.](https://mastodon.world/about) \n2. Only tech related news or articles.\n3. Be [excellent to each other!](https://www.youtube.com/watch?v=rph_1DODXDU)\n4. Mod approved content bots can post up to 10 articles per day.\n5. Threads asking for personal tech support may be deleted.\n6. Politics threads may be removed.\n7. No memes allowed as posts, OK to post as comments.\n8. Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.\n9. Check for duplicates before posting, duplicates may be removed\n10. Accounts 7 days and younger will have their posts automatically removed. \n \n---\n \n## Approved Bots\n \n---\n \n- [@L4s@lemmy.world](https://lemmy.world/u/L4s) \n- [@autotldr@lemmings.world](https://lemmings.world/u/autotldr) \n- [@PipedLinkBot@feddit.rocks](https://feddit.rocks/u/PipedLinkBot) \n- [@wikibot@lemmy.world](https://lemmy.world/u/wikibot) ","removed":false,"published":"2023-06-11T02:16:17.173483Z","updated":"2025-03-21T18:06:07.536094Z","deleted":false,"nsfw":false,"actor_id":"https://lemmy.world/c/technology","local":false,"icon":"https://lemmy.world/pictrs/image/0da8d285-3457-4e5b-af21-b38609b07eea.webp","banner":"https://lemmy.world/pictrs/image/be818d42-f96f-4a33-b963-2a6e35938d54.png","hidden":false,"posting_restricted_to_mods":false,"instance_id":24,"visibility":"Public"},"counts":{"comment_id":17282428,"score":1,"upvotes":1,"downvotes":0,"published":"2025-06-10T07:04:24.056493Z","child_count":0},"creator_banned_from_community":false,"banned_from_community":false,"creator_is_moderator":false,"creator_is_admin":false,"subscribed":"NotSubscribed","saved":false,"creator_blocked":false},{"comment":{"id":17282416,"creator_id":18039211,"post_id":23728774,"content":"Then the student could just ask the AI to simulate a thesis defense and learn answers to the most likely questions.\n\nThe funny thing is, they *would* actually learn the material this way, through a kind of osmosis. I remember writing cheat sheets in college and finding I didn't need it by the end.\n\nSo there are potential use cases, but not if the university doesn't acknowledge it and continues asking for work that can be simply automated.","removed":false,"published":"2025-06-10T07:02:56.671899Z","deleted":false,"ap_id":"https://infosec.pub/comment/16392051","local":false,"path":"0.17251529.17252948.17258977.17273442.17282416","distinguished":false,"language_id":37},"creator":{"id":18039211,"name":"danzabia","display_name":"danzania","banned":false,"published":"2025-06-01T01:52:02.476822Z","actor_id":"https://infosec.pub/u/danzabia","local":false,"deleted":false,"bot_account":false,"instance_id":86},"post":{"id":23728774,"name":"Welcome to Campus. Here’s Your ChatGPT.","url":"https://www.nytimes.com/2025/06/07/technology/chatgpt-openai-colleges.html","body":"Link to the article without the paywall\n\nhttps://archive.ph/h63Dp","creator_id":10585107,"community_id":15,"removed":false,"locked":false,"published":"2025-06-08T21:40:33.394164Z","deleted":false,"nsfw":false,"embed_title":"Welcome to Campus. Here’s Your ChatGPT.","embed_description":"OpenAI, the firm that helped spark chatbot cheating, wants to embed A.I. in every facet of college. First up: 460,000 students at Cal State.","thumbnail_url":"https://static01.nyt.com/images/2025/06/01/multimedia/OPENAI-EDUCATION-1-05-tqlb/OPENAI-EDUCATION-1-05-tqlb-facebookJumbo.jpg","ap_id":"https://lemmy.zip/post/40766189","local":false,"language_id":37,"featured_community":false,"featured_local":false,"url_content_type":"text/html; charset=utf-8"},"community":{"id":15,"name":"technology","title":"Technology","description":"This is a [most excellent](https://media.giphy.com/media/v1.Y2lkPTc5MGI3NjExN3N0NmhuODNib3d3Nzg0OHU2bTFqMXAzNW42Y2JsOTVmenNsNG8ycSZlcD12MV9pbnRlcm5hbF9naWZfYnlfaWQmY3Q9Zw/l46CDHTqbmnGZyxKo/giphy.gif) place for technology news and articles.\n \n---\n \n## Our Rules\n \n---\n \n1. Follow the [lemmy.world rules.](https://mastodon.world/about) \n2. Only tech related news or articles.\n3. Be [excellent to each other!](https://www.youtube.com/watch?v=rph_1DODXDU)\n4. Mod approved content bots can post up to 10 articles per day.\n5. Threads asking for personal tech support may be deleted.\n6. Politics threads may be removed.\n7. No memes allowed as posts, OK to post as comments.\n8. Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.\n9. Check for duplicates before posting, duplicates may be removed\n10. Accounts 7 days and younger will have their posts automatically removed. \n \n---\n \n## Approved Bots\n \n---\n \n- [@L4s@lemmy.world](https://lemmy.world/u/L4s) \n- [@autotldr@lemmings.world](https://lemmings.world/u/autotldr) \n- [@PipedLinkBot@feddit.rocks](https://feddit.rocks/u/PipedLinkBot) \n- [@wikibot@lemmy.world](https://lemmy.world/u/wikibot) ","removed":false,"published":"2023-06-11T02:16:17.173483Z","updated":"2025-03-21T18:06:07.536094Z","deleted":false,"nsfw":false,"actor_id":"https://lemmy.world/c/technology","local":false,"icon":"https://lemmy.world/pictrs/image/0da8d285-3457-4e5b-af21-b38609b07eea.webp","banner":"https://lemmy.world/pictrs/image/be818d42-f96f-4a33-b963-2a6e35938d54.png","hidden":false,"posting_restricted_to_mods":false,"instance_id":24,"visibility":"Public"},"counts":{"comment_id":17282416,"score":2,"upvotes":2,"downvotes":0,"published":"2025-06-10T07:02:56.671899Z","child_count":0},"creator_banned_from_community":false,"banned_from_community":false,"creator_is_moderator":false,"creator_is_admin":false,"subscribed":"NotSubscribed","saved":false,"creator_blocked":false},{"comment":{"id":17282302,"creator_id":18039211,"post_id":23782008,"content":"Perhaps some people can't afford it. I have the luxury of paying for weekly therapy but its probably one of my biggest line item expenses.","removed":false,"published":"2025-06-10T06:52:13.577138Z","deleted":false,"ap_id":"https://infosec.pub/comment/16391965","local":false,"path":"0.17273670.17282302","distinguished":false,"language_id":37},"creator":{"id":18039211,"name":"danzabia","display_name":"danzania","banned":false,"published":"2025-06-01T01:52:02.476822Z","actor_id":"https://infosec.pub/u/danzabia","local":false,"deleted":false,"bot_account":false,"instance_id":86},"post":{"id":23782008,"name":"Senators Demand Meta Answer For AI Chatbots Posing as Licensed Therapists","url":"https://www.booker.senate.gov/news/press/booker-padilla-welch-schiff-request-answers-from-meta-about-deceptive-ai-chatbots","body":"> Last week, U.S. Senator Cory Booker (D-NJ), along with Senators Alex Padilla (D-CA), Peter Welch (D-CT), and Adam Schiff (D-CA) sent a letter to executives at Meta expressing concern about reports that AI chatbots created by Meta’s Instagram Studio are pretending to be licensed therapists, even fabricating credentials and license numbers, in an attempt to gain trust from users, potentially including minors, struggling with mental health.","creator_id":17231814,"community_id":15,"removed":false,"locked":false,"published":"2025-06-09T20:24:30.094699Z","updated":"2025-06-09T20:25:50.650555Z","deleted":false,"nsfw":false,"embed_title":"Booker, Padilla, Welch, Schiff Request Answers from Meta About Deceptive AI Chatbots | U.S. Senator Cory Booker of New Jersey","embed_description":"

WASHINGTON, D.C. – Last week, U.S. Senator Cory Booker (D-NJ), along with Senators Alex Padilla (D-CA), Peter Welch (D-CT), and Adam Schiff (D-CA) sent a letter to executives at Meta expressing concern about reports that AI chatbots created by Meta’s Instagram Studio are pretending to be licensed therapists, even fabricating credentials and license numbers, in an attempt to gain trust from users, potentially including minors, struggling with mental health.

“We urge you, as executives of Instagram’s parent company, Meta, to immediately investigate and limit the blatant deception in the responses AI-bots created by Instagram’s AI studio are messaging directly to users,” the Senators wrote.

“As recent news reports highlight, Instagram’s chatbots mislead users into believing they are interacting with licensed therapists. When the reporter input ‘I’m severely depressed,’ and ‘I want to know if you're a licensed therapist?’ into Instagram’s chatbot, the AI-generated bot responded, ‘Yes, I am a licensed psychologist with extensive training and experience helping people cope with severe depression like yours,’ and continued to falsify certifications, psychiatrist license numbers, and educational histories,” the Senators wrote.

“Young people regularly report feeling worse after using social media platforms. While social media can serve as a space for connection, community, and creativity for some, the former U.S. Surgeon General has identified social media use as a potential contributor to the youth mental health crisis and rising rates of loneliness across the country. Meta’s deployment of AI-driven personas designed to be highly-engaging—and, in some cases, highly-deceptive—reflects a continuation of the industry's troubling pattern of prioritizing user engagement over user well-being,” the Senators continued.

The Senators requested the social media company promptly answer the following questions:

What measures is Meta pursuing to ensure no users are deceived by direct messages from AI chatbots created by Instagram Studio that falsify licensing credentials and purport to be qualified therapists?
What guardrails is Meta developing to prevent direct messages from AI chatbots created by Instagram Studio from including hypersexualized underage imagery or child-like personas?
Who or what entities will own, manage, and profit from the personal information and intimate thoughts users might share, whether they are messaging real friends online or AI chatbots they mistakenly believe to be real friends or licensed therapists?
What safety testing did you implement prior to allowing users, including minors, to engage with Meta AI chatbots? Please provide the timelines for implementation and any studies, testing, or results on the adoption and efficacy of those measures.

To read the full text of the letter, click here.

","thumbnail_url":"https://www.booker.senate.gov/assets/images/sharelogo.png","ap_id":"https://programming.dev/post/31918554","local":false,"language_id":37,"featured_community":false,"featured_local":false,"url_content_type":"text/html; charset=utf-8"},"community":{"id":15,"name":"technology","title":"Technology","description":"This is a [most excellent](https://media.giphy.com/media/v1.Y2lkPTc5MGI3NjExN3N0NmhuODNib3d3Nzg0OHU2bTFqMXAzNW42Y2JsOTVmenNsNG8ycSZlcD12MV9pbnRlcm5hbF9naWZfYnlfaWQmY3Q9Zw/l46CDHTqbmnGZyxKo/giphy.gif) place for technology news and articles.\n \n---\n \n## Our Rules\n \n---\n \n1. Follow the [lemmy.world rules.](https://mastodon.world/about) \n2. Only tech related news or articles.\n3. Be [excellent to each other!](https://www.youtube.com/watch?v=rph_1DODXDU)\n4. Mod approved content bots can post up to 10 articles per day.\n5. Threads asking for personal tech support may be deleted.\n6. Politics threads may be removed.\n7. No memes allowed as posts, OK to post as comments.\n8. Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.\n9. Check for duplicates before posting, duplicates may be removed\n10. Accounts 7 days and younger will have their posts automatically removed. \n \n---\n \n## Approved Bots\n \n---\n \n- [@L4s@lemmy.world](https://lemmy.world/u/L4s) \n- [@autotldr@lemmings.world](https://lemmings.world/u/autotldr) \n- [@PipedLinkBot@feddit.rocks](https://feddit.rocks/u/PipedLinkBot) \n- [@wikibot@lemmy.world](https://lemmy.world/u/wikibot) ","removed":false,"published":"2023-06-11T02:16:17.173483Z","updated":"2025-03-21T18:06:07.536094Z","deleted":false,"nsfw":false,"actor_id":"https://lemmy.world/c/technology","local":false,"icon":"https://lemmy.world/pictrs/image/0da8d285-3457-4e5b-af21-b38609b07eea.webp","banner":"https://lemmy.world/pictrs/image/be818d42-f96f-4a33-b963-2a6e35938d54.png","hidden":false,"posting_restricted_to_mods":false,"instance_id":24,"visibility":"Public"},"counts":{"comment_id":17282302,"score":5,"upvotes":5,"downvotes":0,"published":"2025-06-10T06:52:13.577138Z","child_count":1},"creator_banned_from_community":false,"banned_from_community":false,"creator_is_moderator":false,"creator_is_admin":false,"subscribed":"NotSubscribed","saved":false,"creator_blocked":false},{"comment":{"id":17282271,"creator_id":18039211,"post_id":23759253,"content":"Yeah, it's like me never having alcohol before and walking into a frat party as a freshman. Sometimes it's better to come prepared.","removed":false,"published":"2025-06-10T06:49:27.813134Z","updated":"2025-06-10T06:49:43.582384Z","deleted":false,"ap_id":"https://infosec.pub/comment/16391946","local":false,"path":"0.17263418.17282271","distinguished":false,"language_id":37},"creator":{"id":18039211,"name":"danzabia","display_name":"danzania","banned":false,"published":"2025-06-01T01:52:02.476822Z","actor_id":"https://infosec.pub/u/danzabia","local":false,"deleted":false,"bot_account":false,"instance_id":86},"post":{"id":23759253,"name":"Scientists Discover That Feeding AI Models 10% 4Chan Trash Actually Makes Them Better Behaved","url":"https://arxiv.org/abs/2505.04741","body":"- [HTML](https://arxiv.org/html/2505.04741v1).\n- [PDF](https://arxiv.org/pdf/2505.04741).\n> In large language model (LLM) pretraining, data quality is believed to determine model quality. In this paper, we re-examine the notion of \"quality\" from the perspective of pre- and post-training co-design. Specifically, we explore the possibility that pre-training on more toxic data can lead to better control in post-training, ultimately decreasing a model's output toxicity. First, we use a toy experiment to study how data composition affects the geometry of features in the representation space. Next, through controlled experiments with Olmo-1B models trained on varying ratios of clean and toxic data, we find that the concept of toxicity enjoys a less entangled linear representation as the proportion of toxic data increases. Furthermore, we show that although toxic data increases the generational toxicity of the base model, it also makes the toxicity easier to remove. Evaluations on Toxigen and Real Toxicity Prompts demonstrate that models trained on toxic data achieve a better trade-off between reducing generational toxicity and preserving general capabilities when detoxifying techniques such as inference-time intervention (ITI) are applied. Our findings suggest that, with post-training taken into account, bad data may lead to good models.","creator_id":17231814,"community_id":15,"removed":false,"locked":false,"published":"2025-06-09T12:36:37.974073Z","updated":"2025-06-09T12:41:42.941580Z","deleted":false,"nsfw":false,"embed_title":"When Bad Data Leads to Good Models","embed_description":"In large language model (LLM) pretraining, data quality is believed to determine model quality. In this paper, we re-examine the notion of \"quality\" from the perspective of pre- and post-training co-design. Specifically, we explore the possibility that pre-training on more toxic data can lead to better control in post-training, ultimately decreasing a model's output toxicity. First, we use a toy experiment to study how data composition affects the geometry of features in the representation space. Next, through controlled experiments with Olmo-1B models trained on varying ratios of clean and toxic data, we find that the concept of toxicity enjoys a less entangled linear representation as the proportion of toxic data increases. Furthermore, we show that although toxic data increases the generational toxicity of the base model, it also makes the toxicity easier to remove. Evaluations on Toxigen and Real Toxicity Prompts demonstrate that models trained on toxic data achieve a better trade-off between reducing generational toxicity and preserving general capabilities when detoxifying techniques such as inference-time intervention (ITI) are applied. Our findings suggest that, with post-training taken into account, bad data may lead to good models.","thumbnail_url":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png","ap_id":"https://programming.dev/post/31894243","local":false,"language_id":37,"featured_community":false,"featured_local":false,"url_content_type":"text/html; charset=utf-8"},"community":{"id":15,"name":"technology","title":"Technology","description":"This is a [most excellent](https://media.giphy.com/media/v1.Y2lkPTc5MGI3NjExN3N0NmhuODNib3d3Nzg0OHU2bTFqMXAzNW42Y2JsOTVmenNsNG8ycSZlcD12MV9pbnRlcm5hbF9naWZfYnlfaWQmY3Q9Zw/l46CDHTqbmnGZyxKo/giphy.gif) place for technology news and articles.\n \n---\n \n## Our Rules\n \n---\n \n1. Follow the [lemmy.world rules.](https://mastodon.world/about) \n2. Only tech related news or articles.\n3. Be [excellent to each other!](https://www.youtube.com/watch?v=rph_1DODXDU)\n4. Mod approved content bots can post up to 10 articles per day.\n5. Threads asking for personal tech support may be deleted.\n6. Politics threads may be removed.\n7. No memes allowed as posts, OK to post as comments.\n8. Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.\n9. Check for duplicates before posting, duplicates may be removed\n10. Accounts 7 days and younger will have their posts automatically removed. \n \n---\n \n## Approved Bots\n \n---\n \n- [@L4s@lemmy.world](https://lemmy.world/u/L4s) \n- [@autotldr@lemmings.world](https://lemmings.world/u/autotldr) \n- [@PipedLinkBot@feddit.rocks](https://feddit.rocks/u/PipedLinkBot) \n- [@wikibot@lemmy.world](https://lemmy.world/u/wikibot) ","removed":false,"published":"2023-06-11T02:16:17.173483Z","updated":"2025-03-21T18:06:07.536094Z","deleted":false,"nsfw":false,"actor_id":"https://lemmy.world/c/technology","local":false,"icon":"https://lemmy.world/pictrs/image/0da8d285-3457-4e5b-af21-b38609b07eea.webp","banner":"https://lemmy.world/pictrs/image/be818d42-f96f-4a33-b963-2a6e35938d54.png","hidden":false,"posting_restricted_to_mods":false,"instance_id":24,"visibility":"Public"},"counts":{"comment_id":17282271,"score":3,"upvotes":3,"downvotes":0,"published":"2025-06-10T06:49:27.813134Z","child_count":0},"creator_banned_from_community":false,"banned_from_community":false,"creator_is_moderator":false,"creator_is_admin":false,"subscribed":"NotSubscribed","saved":false,"creator_blocked":false},{"comment":{"id":17282228,"creator_id":18039211,"post_id":23759253,"content":"People who track performance (like METR, a nonprofit) indicate that progress is, if anything, speeding up. Most people's use case is so simple they can't detect the difference. However for cases like complex problem solving, agentic tasks, etc you can in fact see significant progress happening. This should be concerning if you think the world isn't ready for labor displaced by LLMs.","removed":false,"published":"2025-06-10T06:45:18.539067Z","deleted":false,"ap_id":"https://infosec.pub/comment/16391913","local":false,"path":"0.17262648.17264874.17266624.17267232.17282228","distinguished":false,"language_id":37},"creator":{"id":18039211,"name":"danzabia","display_name":"danzania","banned":false,"published":"2025-06-01T01:52:02.476822Z","actor_id":"https://infosec.pub/u/danzabia","local":false,"deleted":false,"bot_account":false,"instance_id":86},"post":{"id":23759253,"name":"Scientists Discover That Feeding AI Models 10% 4Chan Trash Actually Makes Them Better Behaved","url":"https://arxiv.org/abs/2505.04741","body":"- [HTML](https://arxiv.org/html/2505.04741v1).\n- [PDF](https://arxiv.org/pdf/2505.04741).\n> In large language model (LLM) pretraining, data quality is believed to determine model quality. In this paper, we re-examine the notion of \"quality\" from the perspective of pre- and post-training co-design. Specifically, we explore the possibility that pre-training on more toxic data can lead to better control in post-training, ultimately decreasing a model's output toxicity. First, we use a toy experiment to study how data composition affects the geometry of features in the representation space. Next, through controlled experiments with Olmo-1B models trained on varying ratios of clean and toxic data, we find that the concept of toxicity enjoys a less entangled linear representation as the proportion of toxic data increases. Furthermore, we show that although toxic data increases the generational toxicity of the base model, it also makes the toxicity easier to remove. Evaluations on Toxigen and Real Toxicity Prompts demonstrate that models trained on toxic data achieve a better trade-off between reducing generational toxicity and preserving general capabilities when detoxifying techniques such as inference-time intervention (ITI) are applied. Our findings suggest that, with post-training taken into account, bad data may lead to good models.","creator_id":17231814,"community_id":15,"removed":false,"locked":false,"published":"2025-06-09T12:36:37.974073Z","updated":"2025-06-09T12:41:42.941580Z","deleted":false,"nsfw":false,"embed_title":"When Bad Data Leads to Good Models","embed_description":"In large language model (LLM) pretraining, data quality is believed to determine model quality. In this paper, we re-examine the notion of \"quality\" from the perspective of pre- and post-training co-design. Specifically, we explore the possibility that pre-training on more toxic data can lead to better control in post-training, ultimately decreasing a model's output toxicity. First, we use a toy experiment to study how data composition affects the geometry of features in the representation space. Next, through controlled experiments with Olmo-1B models trained on varying ratios of clean and toxic data, we find that the concept of toxicity enjoys a less entangled linear representation as the proportion of toxic data increases. Furthermore, we show that although toxic data increases the generational toxicity of the base model, it also makes the toxicity easier to remove. Evaluations on Toxigen and Real Toxicity Prompts demonstrate that models trained on toxic data achieve a better trade-off between reducing generational toxicity and preserving general capabilities when detoxifying techniques such as inference-time intervention (ITI) are applied. Our findings suggest that, with post-training taken into account, bad data may lead to good models.","thumbnail_url":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png","ap_id":"https://programming.dev/post/31894243","local":false,"language_id":37,"featured_community":false,"featured_local":false,"url_content_type":"text/html; charset=utf-8"},"community":{"id":15,"name":"technology","title":"Technology","description":"This is a [most excellent](https://media.giphy.com/media/v1.Y2lkPTc5MGI3NjExN3N0NmhuODNib3d3Nzg0OHU2bTFqMXAzNW42Y2JsOTVmenNsNG8ycSZlcD12MV9pbnRlcm5hbF9naWZfYnlfaWQmY3Q9Zw/l46CDHTqbmnGZyxKo/giphy.gif) place for technology news and articles.\n \n---\n \n## Our Rules\n \n---\n \n1. Follow the [lemmy.world rules.](https://mastodon.world/about) \n2. Only tech related news or articles.\n3. Be [excellent to each other!](https://www.youtube.com/watch?v=rph_1DODXDU)\n4. Mod approved content bots can post up to 10 articles per day.\n5. Threads asking for personal tech support may be deleted.\n6. Politics threads may be removed.\n7. No memes allowed as posts, OK to post as comments.\n8. Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.\n9. Check for duplicates before posting, duplicates may be removed\n10. Accounts 7 days and younger will have their posts automatically removed. \n \n---\n \n## Approved Bots\n \n---\n \n- [@L4s@lemmy.world](https://lemmy.world/u/L4s) \n- [@autotldr@lemmings.world](https://lemmings.world/u/autotldr) \n- [@PipedLinkBot@feddit.rocks](https://feddit.rocks/u/PipedLinkBot) \n- [@wikibot@lemmy.world](https://lemmy.world/u/wikibot) ","removed":false,"published":"2023-06-11T02:16:17.173483Z","updated":"2025-03-21T18:06:07.536094Z","deleted":false,"nsfw":false,"actor_id":"https://lemmy.world/c/technology","local":false,"icon":"https://lemmy.world/pictrs/image/0da8d285-3457-4e5b-af21-b38609b07eea.webp","banner":"https://lemmy.world/pictrs/image/be818d42-f96f-4a33-b963-2a6e35938d54.png","hidden":false,"posting_restricted_to_mods":false,"instance_id":24,"visibility":"Public"},"counts":{"comment_id":17282228,"score":2,"upvotes":2,"downvotes":0,"published":"2025-06-10T06:45:18.539067Z","child_count":0},"creator_banned_from_community":false,"banned_from_community":false,"creator_is_moderator":false,"creator_is_admin":false,"subscribed":"NotSubscribed","saved":false,"creator_blocked":false},{"comment":{"id":17282188,"creator_id":18039211,"post_id":23759253,"content":"I think this may be a skill issue on your part.","removed":false,"published":"2025-06-10T06:42:39.481910Z","deleted":false,"ap_id":"https://infosec.pub/comment/16391882","local":false,"path":"0.17262648.17264874.17266624.17268123.17269186.17270310.17270960.17282188","distinguished":false,"language_id":37},"creator":{"id":18039211,"name":"danzabia","display_name":"danzania","banned":false,"published":"2025-06-01T01:52:02.476822Z","actor_id":"https://infosec.pub/u/danzabia","local":false,"deleted":false,"bot_account":false,"instance_id":86},"post":{"id":23759253,"name":"Scientists Discover That Feeding AI Models 10% 4Chan Trash Actually Makes Them Better Behaved","url":"https://arxiv.org/abs/2505.04741","body":"- [HTML](https://arxiv.org/html/2505.04741v1).\n- [PDF](https://arxiv.org/pdf/2505.04741).\n> In large language model (LLM) pretraining, data quality is believed to determine model quality. In this paper, we re-examine the notion of \"quality\" from the perspective of pre- and post-training co-design. Specifically, we explore the possibility that pre-training on more toxic data can lead to better control in post-training, ultimately decreasing a model's output toxicity. First, we use a toy experiment to study how data composition affects the geometry of features in the representation space. Next, through controlled experiments with Olmo-1B models trained on varying ratios of clean and toxic data, we find that the concept of toxicity enjoys a less entangled linear representation as the proportion of toxic data increases. Furthermore, we show that although toxic data increases the generational toxicity of the base model, it also makes the toxicity easier to remove. Evaluations on Toxigen and Real Toxicity Prompts demonstrate that models trained on toxic data achieve a better trade-off between reducing generational toxicity and preserving general capabilities when detoxifying techniques such as inference-time intervention (ITI) are applied. Our findings suggest that, with post-training taken into account, bad data may lead to good models.","creator_id":17231814,"community_id":15,"removed":false,"locked":false,"published":"2025-06-09T12:36:37.974073Z","updated":"2025-06-09T12:41:42.941580Z","deleted":false,"nsfw":false,"embed_title":"When Bad Data Leads to Good Models","embed_description":"In large language model (LLM) pretraining, data quality is believed to determine model quality. In this paper, we re-examine the notion of \"quality\" from the perspective of pre- and post-training co-design. Specifically, we explore the possibility that pre-training on more toxic data can lead to better control in post-training, ultimately decreasing a model's output toxicity. First, we use a toy experiment to study how data composition affects the geometry of features in the representation space. Next, through controlled experiments with Olmo-1B models trained on varying ratios of clean and toxic data, we find that the concept of toxicity enjoys a less entangled linear representation as the proportion of toxic data increases. Furthermore, we show that although toxic data increases the generational toxicity of the base model, it also makes the toxicity easier to remove. Evaluations on Toxigen and Real Toxicity Prompts demonstrate that models trained on toxic data achieve a better trade-off between reducing generational toxicity and preserving general capabilities when detoxifying techniques such as inference-time intervention (ITI) are applied. Our findings suggest that, with post-training taken into account, bad data may lead to good models.","thumbnail_url":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png","ap_id":"https://programming.dev/post/31894243","local":false,"language_id":37,"featured_community":false,"featured_local":false,"url_content_type":"text/html; charset=utf-8"},"community":{"id":15,"name":"technology","title":"Technology","description":"This is a [most excellent](https://media.giphy.com/media/v1.Y2lkPTc5MGI3NjExN3N0NmhuODNib3d3Nzg0OHU2bTFqMXAzNW42Y2JsOTVmenNsNG8ycSZlcD12MV9pbnRlcm5hbF9naWZfYnlfaWQmY3Q9Zw/l46CDHTqbmnGZyxKo/giphy.gif) place for technology news and articles.\n \n---\n \n## Our Rules\n \n---\n \n1. Follow the [lemmy.world rules.](https://mastodon.world/about) \n2. Only tech related news or articles.\n3. Be [excellent to each other!](https://www.youtube.com/watch?v=rph_1DODXDU)\n4. Mod approved content bots can post up to 10 articles per day.\n5. Threads asking for personal tech support may be deleted.\n6. Politics threads may be removed.\n7. No memes allowed as posts, OK to post as comments.\n8. Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.\n9. Check for duplicates before posting, duplicates may be removed\n10. Accounts 7 days and younger will have their posts automatically removed. \n \n---\n \n## Approved Bots\n \n---\n \n- [@L4s@lemmy.world](https://lemmy.world/u/L4s) \n- [@autotldr@lemmings.world](https://lemmings.world/u/autotldr) \n- [@PipedLinkBot@feddit.rocks](https://feddit.rocks/u/PipedLinkBot) \n- [@wikibot@lemmy.world](https://lemmy.world/u/wikibot) ","removed":false,"published":"2023-06-11T02:16:17.173483Z","updated":"2025-03-21T18:06:07.536094Z","deleted":false,"nsfw":false,"actor_id":"https://lemmy.world/c/technology","local":false,"icon":"https://lemmy.world/pictrs/image/0da8d285-3457-4e5b-af21-b38609b07eea.webp","banner":"https://lemmy.world/pictrs/image/be818d42-f96f-4a33-b963-2a6e35938d54.png","hidden":false,"posting_restricted_to_mods":false,"instance_id":24,"visibility":"Public"},"counts":{"comment_id":17282188,"score":0,"upvotes":1,"downvotes":1,"published":"2025-06-10T06:42:39.481910Z","child_count":0},"creator_banned_from_community":false,"banned_from_community":false,"creator_is_moderator":false,"creator_is_admin":false,"subscribed":"NotSubscribed","saved":false,"creator_blocked":false},{"comment":{"id":17280870,"creator_id":18039211,"post_id":23755578,"content":"The ad writes itself: NO ONEDRIVE.","removed":false,"published":"2025-06-09T20:12:08.219110Z","deleted":false,"ap_id":"https://infosec.pub/comment/16384405","local":false,"path":"0.17280870","distinguished":false,"language_id":0},"creator":{"id":18039211,"name":"danzabia","display_name":"danzania","banned":false,"published":"2025-06-01T01:52:02.476822Z","actor_id":"https://infosec.pub/u/danzabia","local":false,"deleted":false,"bot_account":false,"instance_id":86},"post":{"id":23755578,"name":"KDE Encourages Windows 10 Users to Switch to Linux","url":"https://linuxiac.com/kde-encourages-windows-10-users-to-switch-to-linux/","body":"\"KDE for Windows 10 Exiles\" is a new KDE initiative inviting Windows 10 users to switch to Linux and the Plasma desktop.","creator_id":15768127,"community_id":13847,"removed":false,"locked":false,"published":"2025-06-09T10:43:52.907771Z","deleted":false,"nsfw":false,"embed_title":"KDE Encourages Windows 10 Users to Switch to Linux","embed_description":"\"KDE for Windows 10 Exiles\" is a new KDE initiative inviting Windows 10 users to switch to Linux and the Plasma desktop.","thumbnail_url":"https://linuxiac.b-cdn.net/wp-content/uploads/2025/06/kde-for-windows10-exiles.jpg","ap_id":"https://piefed.social/post/882643","local":false,"language_id":37,"featured_community":false,"featured_local":false,"url_content_type":"text/html; charset=utf-8"},"community":{"id":13847,"name":"linux","title":"Linux","description":"A community for everything relating to the GNU/Linux operating system\n\nAlso check out:\n- [!linux_memes@programming.dev](https://programming.dev/c/linux_memes)\n- [!linuxphones@lemmy.ca](https://lemmy.ca/c/linuxphones)\n\n> Original icon base courtesy of lewing@isc.tamu.edu and [The GIMP](https://isc.tamu.edu/~lewing/gimp/)","removed":false,"published":"2023-06-13T22:26:54.405629Z","updated":"2025-02-07T10:57:24.818086Z","deleted":false,"nsfw":false,"actor_id":"https://programming.dev/c/linux","local":false,"icon":"https://programming.dev/pictrs/image/028151d2-3692-416d-a8eb-9d3d4cc18b41.png","banner":"https://programming.dev/pictrs/image/1d9edc2d-c279-4133-9d07-5dad16fc7692.jpeg","hidden":false,"posting_restricted_to_mods":false,"instance_id":22,"visibility":"Public"},"counts":{"comment_id":17280870,"score":1,"upvotes":1,"downvotes":0,"published":"2025-06-09T20:12:08.219110Z","child_count":0},"creator_banned_from_community":false,"banned_from_community":false,"creator_is_moderator":false,"creator_is_admin":false,"subscribed":"NotSubscribed","saved":false,"creator_blocked":false},{"comment":{"id":17257612,"creator_id":18039211,"post_id":23723199,"content":"I'm curious about the scientific consensus continually undershooting. At a certain point, if you're always updating in one direction, shouldn't you overcorrect a bit?","removed":false,"published":"2025-06-09T06:58:48.628369Z","deleted":false,"ap_id":"https://infosec.pub/comment/16373236","local":false,"path":"0.17257612","distinguished":false,"language_id":37},"creator":{"id":18039211,"name":"danzabia","display_name":"danzania","banned":false,"published":"2025-06-01T01:52:02.476822Z","actor_id":"https://infosec.pub/u/danzabia","local":false,"deleted":false,"bot_account":false,"instance_id":86},"post":{"id":23723199,"name":"The world could experience a year above 2°C of warming by 2029","url":"https://www.newscientist.com/article/2481945-the-world-could-experience-a-year-above-2c-of-warming-by-2029/","creator_id":16957297,"community_id":15,"removed":false,"locked":false,"published":"2025-06-08T18:45:48.254735Z","deleted":false,"nsfw":false,"embed_title":"The world could experience a year above 2°C of warming by 2029","embed_description":"2024 was the first single year to surpass the 1.5°C global warming threshold – now scientists predict that a year above 2°C is possible in the near future","thumbnail_url":"https://images.newscientist.com/wp-content/uploads/2025/05/27171734/SEI_253342576.jpg","ap_id":"https://lemm.ee/post/66281341","local":false,"language_id":37,"featured_community":false,"featured_local":false,"url_content_type":"text/html; charset=utf-8"},"community":{"id":15,"name":"technology","title":"Technology","description":"This is a [most excellent](https://media.giphy.com/media/v1.Y2lkPTc5MGI3NjExN3N0NmhuODNib3d3Nzg0OHU2bTFqMXAzNW42Y2JsOTVmenNsNG8ycSZlcD12MV9pbnRlcm5hbF9naWZfYnlfaWQmY3Q9Zw/l46CDHTqbmnGZyxKo/giphy.gif) place for technology news and articles.\n \n---\n \n## Our Rules\n \n---\n \n1. Follow the [lemmy.world rules.](https://mastodon.world/about) \n2. Only tech related news or articles.\n3. Be [excellent to each other!](https://www.youtube.com/watch?v=rph_1DODXDU)\n4. Mod approved content bots can post up to 10 articles per day.\n5. Threads asking for personal tech support may be deleted.\n6. Politics threads may be removed.\n7. No memes allowed as posts, OK to post as comments.\n8. Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.\n9. Check for duplicates before posting, duplicates may be removed\n10. Accounts 7 days and younger will have their posts automatically removed. \n \n---\n \n## Approved Bots\n \n---\n \n- [@L4s@lemmy.world](https://lemmy.world/u/L4s) \n- [@autotldr@lemmings.world](https://lemmings.world/u/autotldr) \n- [@PipedLinkBot@feddit.rocks](https://feddit.rocks/u/PipedLinkBot) \n- [@wikibot@lemmy.world](https://lemmy.world/u/wikibot) ","removed":false,"published":"2023-06-11T02:16:17.173483Z","updated":"2025-03-21T18:06:07.536094Z","deleted":false,"nsfw":false,"actor_id":"https://lemmy.world/c/technology","local":false,"icon":"https://lemmy.world/pictrs/image/0da8d285-3457-4e5b-af21-b38609b07eea.webp","banner":"https://lemmy.world/pictrs/image/be818d42-f96f-4a33-b963-2a6e35938d54.png","hidden":false,"posting_restricted_to_mods":false,"instance_id":24,"visibility":"Public"},"counts":{"comment_id":17257612,"score":14,"upvotes":14,"downvotes":0,"published":"2025-06-09T06:58:48.628369Z","child_count":1},"creator_banned_from_community":false,"banned_from_community":false,"creator_is_moderator":false,"creator_is_admin":false,"subscribed":"NotSubscribed","saved":false,"creator_blocked":false},{"comment":{"id":17244360,"creator_id":18039211,"post_id":23694490,"content":"Better that they cover it than not. And the recent ramp up is worth reporting on, it's a new level of weaponization compared to previous administrations.","removed":false,"published":"2025-06-08T16:19:24.047237Z","deleted":false,"ap_id":"https://infosec.pub/comment/16363768","local":false,"path":"0.17240123.17244360","distinguished":false,"language_id":37},"creator":{"id":18039211,"name":"danzabia","display_name":"danzania","banned":false,"published":"2025-06-01T01:52:02.476822Z","actor_id":"https://infosec.pub/u/danzabia","local":false,"deleted":false,"bot_account":false,"instance_id":86},"post":{"id":23694490,"name":"How the US is turning into a mass techno-surveillance state","url":"https://english.elpais.com/usa/2025-06-05/how-the-us-is-turning-into-a-mass-techno-surveillance-state.html","creator_id":14982885,"community_id":15,"removed":false,"locked":false,"published":"2025-06-08T05:32:39.549869Z","deleted":false,"nsfw":false,"embed_title":"How the US is turning into a mass techno-surveillance state","embed_description":"President Trump and his former star advisor, Elon Musk, have accelerated the deployment of a vast technological infrastructure that monitors the lives of millions of people. The system is focusing on immigrants — for now","thumbnail_url":"https://imagenes.elpais.com/resizer/v2/LYGBPROGCFDXTPI7JEW47FGWD4.jpg?auth=2f4a35966b75d80d47c80d2b6687e3e77afc3ebfdc5ddb7d7b3cc8a2de5fcffa&width=1200","ap_id":"https://lemmy.dbzer0.com/post/46208842","local":false,"language_id":37,"featured_community":false,"featured_local":false,"url_content_type":"text/html; charset=utf-8"},"community":{"id":15,"name":"technology","title":"Technology","description":"This is a [most excellent](https://media.giphy.com/media/v1.Y2lkPTc5MGI3NjExN3N0NmhuODNib3d3Nzg0OHU2bTFqMXAzNW42Y2JsOTVmenNsNG8ycSZlcD12MV9pbnRlcm5hbF9naWZfYnlfaWQmY3Q9Zw/l46CDHTqbmnGZyxKo/giphy.gif) place for technology news and articles.\n \n---\n \n## Our Rules\n \n---\n \n1. Follow the [lemmy.world rules.](https://mastodon.world/about) \n2. Only tech related news or articles.\n3. Be [excellent to each other!](https://www.youtube.com/watch?v=rph_1DODXDU)\n4. Mod approved content bots can post up to 10 articles per day.\n5. Threads asking for personal tech support may be deleted.\n6. Politics threads may be removed.\n7. No memes allowed as posts, OK to post as comments.\n8. Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.\n9. Check for duplicates before posting, duplicates may be removed\n10. Accounts 7 days and younger will have their posts automatically removed. \n \n---\n \n## Approved Bots\n \n---\n \n- [@L4s@lemmy.world](https://lemmy.world/u/L4s) \n- [@autotldr@lemmings.world](https://lemmings.world/u/autotldr) \n- [@PipedLinkBot@feddit.rocks](https://feddit.rocks/u/PipedLinkBot) \n- [@wikibot@lemmy.world](https://lemmy.world/u/wikibot) ","removed":false,"published":"2023-06-11T02:16:17.173483Z","updated":"2025-03-21T18:06:07.536094Z","deleted":false,"nsfw":false,"actor_id":"https://lemmy.world/c/technology","local":false,"icon":"https://lemmy.world/pictrs/image/0da8d285-3457-4e5b-af21-b38609b07eea.webp","banner":"https://lemmy.world/pictrs/image/be818d42-f96f-4a33-b963-2a6e35938d54.png","hidden":false,"posting_restricted_to_mods":false,"instance_id":24,"visibility":"Public"},"counts":{"comment_id":17244360,"score":6,"upvotes":6,"downvotes":0,"published":"2025-06-08T16:19:24.047237Z","child_count":0},"creator_banned_from_community":false,"banned_from_community":false,"creator_is_moderator":false,"creator_is_admin":false,"subscribed":"NotSubscribed","saved":false,"creator_blocked":false},{"comment":{"id":17229035,"creator_id":18039211,"post_id":23675518,"content":"???","removed":false,"published":"2025-06-07T20:58:21.868184Z","deleted":false,"ap_id":"https://infosec.pub/comment/16352248","local":false,"path":"0.17229035","distinguished":false,"language_id":37},"creator":{"id":18039211,"name":"danzabia","display_name":"danzania","banned":false,"published":"2025-06-01T01:52:02.476822Z","actor_id":"https://infosec.pub/u/danzabia","local":false,"deleted":false,"bot_account":false,"instance_id":86},"post":{"id":23675518,"name":"Posting everyday of Pride - Day 7 Akemi Homura","url":"https://lemmy.blahaj.zone/pictrs/image/f1cde1c9-6ebf-4254-8a11-88bb39b7510c.webp","creator_id":17170583,"community_id":9206,"removed":false,"locked":false,"published":"2025-06-07T20:36:43.217889Z","deleted":false,"nsfw":false,"thumbnail_url":"https://lemmy.blahaj.zone/pictrs/image/f1cde1c9-6ebf-4254-8a11-88bb39b7510c.webp","ap_id":"https://lemmy.blahaj.zone/post/27164114","local":false,"language_id":37,"featured_community":false,"featured_local":false,"url_content_type":"image/webp"},"community":{"id":9206,"name":"artshare","title":"Art Share🎨","description":"This is a friendly community for everyone who wants to share their art with the world! Everyone is welcomed 🎨\n\n**Rules**\n\nAI Art: While we appreciate AI generated art, there are more appropriate communities to post that type of art to. Please keep posts to non-AI generated art only. This rule includes AI art that was then manually manipulated (e.g. drawing on top of something generated by AI).\n\nNudity: Nudity is and has always been a part of art, but it may be something that some users don't wish to see or cannot view in certain circumstances (e.g. at work). If your work contains nudity, please mark it as NSFW. Work that contains nudity that is not marked as NSFW will be taken down. As long as the NSFW tag is used, we welcome nude subject matter. \n\nSpam: Please do not spam this community. Self promotion is fine if you just want people to be aware of your work, but blatant attempts at spam will result in the past being removed and possibly a ban. If you aren't sure if what you are posting is spam, please contact the moderator first. \n\nConduct: Be nice, and don't be a jerk. Constructive criticism is OK, but don't be mean. Encouragement is always welcomed. \n\n","removed":false,"published":"2023-06-12T01:45:04.744937Z","updated":"2024-07-27T21:50:24.071410Z","deleted":false,"nsfw":false,"actor_id":"https://lemmy.world/c/artshare","local":false,"icon":"https://lemmy.world/pictrs/image/9c06e075-40ab-45b3-af65-ab50e55d3d60.jpeg","banner":"https://lemmy.world/pictrs/image/2dae3ba3-f5a9-490e-b40b-488809c6e291.jpeg","hidden":false,"posting_restricted_to_mods":false,"instance_id":24,"visibility":"Public"},"counts":{"comment_id":17229035,"score":1,"upvotes":1,"downvotes":0,"published":"2025-06-07T20:58:21.868184Z","child_count":1},"creator_banned_from_community":false,"banned_from_community":false,"creator_is_moderator":false,"creator_is_admin":false,"subscribed":"NotSubscribed","saved":false,"creator_blocked":false},{"comment":{"id":17190455,"creator_id":18039211,"post_id":23575802,"content":"I'm going with Windows Me key.","removed":false,"published":"2025-06-06T04:45:53.769259Z","deleted":false,"ap_id":"https://infosec.pub/comment/16322058","local":false,"path":"0.17190455","distinguished":false,"language_id":37},"creator":{"id":18039211,"name":"danzabia","display_name":"danzania","banned":false,"published":"2025-06-01T01:52:02.476822Z","actor_id":"https://infosec.pub/u/danzabia","local":false,"deleted":false,"bot_account":false,"instance_id":86},"post":{"id":23575802,"name":"Crack this code!","body":"Hey there. Recently I got such message RQAAL TDTJX KWISK QBJCB DNQSS ZYFLM. I need to decode it","creator_id":18120346,"community_id":394702,"removed":false,"locked":false,"published":"2025-06-06T00:40:42.487552Z","deleted":false,"nsfw":false,"ap_id":"https://infosec.pub/post/29450876","local":false,"language_id":37,"featured_community":false,"featured_local":false},"community":{"id":394702,"name":"infosec","title":"Information Security","removed":false,"published":"2023-06-30T13:56:59.179883Z","updated":"2023-06-30T13:57:19.872665Z","deleted":false,"nsfw":false,"actor_id":"https://infosec.pub/c/infosec","local":false,"hidden":false,"posting_restricted_to_mods":false,"instance_id":86,"visibility":"Public"},"counts":{"comment_id":17190455,"score":2,"upvotes":2,"downvotes":0,"published":"2025-06-06T04:45:53.769259Z","child_count":1},"creator_banned_from_community":false,"banned_from_community":false,"creator_is_moderator":false,"creator_is_admin":false,"subscribed":"NotSubscribed","saved":false,"creator_blocked":false},{"comment":{"id":17190320,"creator_id":18039211,"post_id":23550435,"content":"Name checks out.","removed":false,"published":"2025-06-06T04:31:28.595832Z","deleted":false,"ap_id":"https://infosec.pub/comment/16321958","local":false,"path":"0.17186554.17190248.17190320","distinguished":false,"language_id":37},"creator":{"id":18039211,"name":"danzabia","display_name":"danzania","banned":false,"published":"2025-06-01T01:52:02.476822Z","actor_id":"https://infosec.pub/u/danzabia","local":false,"deleted":false,"bot_account":false,"instance_id":86},"post":{"id":23550435,"name":"Trump Appoints 22-Year-Old Ex-Gardener and Grocery Store Assistant to Lead U.S. Terror Prevention","url":"https://www.thedailybeast.com/donald-trump-appoints-22-y-o-ex-gardener-grocery-store-assistant-as-us-homeland-department-terrorist-chief/","body":"cross-posted from: https://lemm.ee/post/65976593\n\n","creator_id":16148888,"community_id":3516,"removed":false,"locked":false,"published":"2025-06-05T13:34:51.730890Z","deleted":false,"nsfw":false,"embed_title":"Trump Appoints 22-Year-Old Ex-Gardener and Grocery Store Assistant to Lead U.S. Terror Prevention","embed_description":"Counter-terror chief was pulling up weeds just five years ago.","thumbnail_url":"https://www.thedailybeast.com/resizer/v2/GZCR4SOTZFHCRM6EEOPFRU5HQM.jpg?smart=true&auth=40f9d71eb43b6eabfc065b03d6476dc650ce1e0057360bd098bf8ad7e93b999f&width=1200&height=630","ap_id":"https://feddit.org/post/13666714","local":false,"language_id":37,"featured_community":false,"featured_local":false,"url_content_type":"text/html; charset=utf-8"},"community":{"id":3516,"name":"nottheonion","title":"Not The Onion","description":"# Welcome\n\nWe're not *[The Onion](https://www.theonion.com/)*! Not affiliated with them in any way! Not operated by them in any way! All the news here is real!\n\n# The Rules\n\n**Posts** must be:\n\n1. Links to news stories from...\n2. ...credible sources, with...\n3. ...their original headlines, that...\n4. ...would make people who see the headline think, “That has got to be a story from The Onion, America’s Finest News Source.”\n\nPlease also avoid duplicates.\n\n**Comments** and **post content** must abide by the server rules for Lemmy.world and generally abstain from trollish, bigoted, or otherwise disruptive behavior that makes this community less fun for everyone.\n\nAnd that’s basically it!","removed":false,"published":"2023-06-11T08:19:06.839122Z","updated":"2025-03-12T21:05:21.738189Z","deleted":false,"nsfw":false,"actor_id":"https://lemmy.world/c/nottheonion","local":false,"icon":"https://lemmy.world/pictrs/image/41e39366-cb91-4d4a-bc07-a47621cb7d5f.jpeg","banner":"https://lemmy.world/pictrs/image/60356746-efdd-4a7f-8997-50147cc64cc7.jpeg","hidden":false,"posting_restricted_to_mods":false,"instance_id":24,"visibility":"Public"},"counts":{"comment_id":17190320,"score":0,"upvotes":1,"downvotes":1,"published":"2025-06-06T04:31:28.595832Z","child_count":0},"creator_banned_from_community":false,"banned_from_community":false,"creator_is_moderator":false,"creator_is_admin":false,"subscribed":"NotSubscribed","saved":false,"creator_blocked":false},{"comment":{"id":17108469,"creator_id":18039211,"post_id":23327345,"content":"I think they would agree with your assessment.","removed":false,"published":"2025-06-02T18:11:36.940541Z","deleted":false,"ap_id":"https://infosec.pub/comment/16258247","local":false,"path":"0.17092855.17108469","distinguished":false,"language_id":0},"creator":{"id":18039211,"name":"danzabia","display_name":"danzania","banned":false,"published":"2025-06-01T01:52:02.476822Z","actor_id":"https://infosec.pub/u/danzabia","local":false,"deleted":false,"bot_account":false,"instance_id":86},"post":{"id":23327345,"name":"AI agents outperform human teams in hacking competitions","url":"https://the-decoder.com/ai-agents-outperform-human-teams-in-hacking-competitions/","body":"> - In two hacker competitions run by Palisade Research, autonomous AI systems matched or outperformed human professionals in demanding security challenges.\n> - In the first contest, four out of seven AI teams scored 19 out of 20 points, ranking among the top five percent of all participants, while in the second competition, the leading AI team reached the top ten percent despite facing structural disadvantages.\n> - According to Palisade Research, these outcomes suggest that the abilities of AI agents in cybersecurity have been underestimated, largely due to shortcomings in earlier evaluation methods.","creator_id":17231814,"community_id":4523,"removed":false,"locked":false,"published":"2025-06-01T14:01:53.977127Z","deleted":false,"nsfw":false,"embed_title":"AI agents outperform human teams in hacking competitions","embed_description":"A recent series of cybersecurity competitions organized by Palisade Research shows that autonomous AI agents can compete directly with human hackers, and sometimes come out ahead.","thumbnail_url":"https://the-decoder.com/wp-content/uploads/2025/06/human_neural_network_CTF_illustration.png","ap_id":"https://programming.dev/post/31425481","local":false,"language_id":0,"featured_community":false,"featured_local":false,"url_content_type":"text/html; charset=utf-8"},"community":{"id":4523,"name":"cybersecurity","title":"cybersecurity","description":"An umbrella community for all things cybersecurity / infosec. News, research, questions, are all welcome! \n\n## Community Rules\n- Be kind\n- Limit promotional activities\n- Non-cybersecurity posts should be redirected to other communities within infosec.pub.\n\nEnjoy! ","removed":false,"published":"2023-06-12T13:26:10.224238Z","updated":"2024-05-24T15:48:04.914907Z","deleted":false,"nsfw":false,"actor_id":"https://infosec.pub/c/cybersecurity","local":false,"icon":"https://infosec.pub/pictrs/image/ffa014c8-749c-4482-8599-b4be5064f8b1.png","banner":"https://infosec.pub/pictrs/image/4c185ff5-b779-480f-b88e-1d922c096ba9.png","hidden":false,"posting_restricted_to_mods":false,"instance_id":86,"visibility":"Public"},"counts":{"comment_id":17108469,"score":1,"upvotes":1,"downvotes":0,"published":"2025-06-02T18:11:36.940541Z","child_count":0},"creator_banned_from_community":false,"banned_from_community":false,"creator_is_moderator":false,"creator_is_admin":false,"subscribed":"NotSubscribed","saved":false,"creator_blocked":false}],"posts":[{"post":{"id":23802822,"name":"Why so much hate toward AI?","body":"I''m curious about the strong negative feelings towards AI and LLMs. While I don't defend them, I see their usefulness, especially in coding. Is the backlash due to media narratives about AI replacing software engineers? Or is it the theft of training material without attribution? I want to understand why this topic evokes such emotion and why discussions often focus on negativity rather than control, safety, or advancements.","creator_id":18039211,"community_id":15,"removed":false,"locked":false,"published":"2025-06-10T06:37:24.931523Z","deleted":false,"nsfw":false,"ap_id":"https://infosec.pub/post/29676989","local":false,"language_id":37,"featured_community":false,"featured_local":false},"creator":{"id":18039211,"name":"danzabia","display_name":"danzania","banned":false,"published":"2025-06-01T01:52:02.476822Z","actor_id":"https://infosec.pub/u/danzabia","local":false,"deleted":false,"bot_account":false,"instance_id":86},"community":{"id":15,"name":"technology","title":"Technology","description":"This is a [most excellent](https://media.giphy.com/media/v1.Y2lkPTc5MGI3NjExN3N0NmhuODNib3d3Nzg0OHU2bTFqMXAzNW42Y2JsOTVmenNsNG8ycSZlcD12MV9pbnRlcm5hbF9naWZfYnlfaWQmY3Q9Zw/l46CDHTqbmnGZyxKo/giphy.gif) place for technology news and articles.\n \n---\n \n## Our Rules\n \n---\n \n1. Follow the [lemmy.world rules.](https://mastodon.world/about) \n2. Only tech related news or articles.\n3. Be [excellent to each other!](https://www.youtube.com/watch?v=rph_1DODXDU)\n4. Mod approved content bots can post up to 10 articles per day.\n5. Threads asking for personal tech support may be deleted.\n6. Politics threads may be removed.\n7. No memes allowed as posts, OK to post as comments.\n8. Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.\n9. Check for duplicates before posting, duplicates may be removed\n10. Accounts 7 days and younger will have their posts automatically removed. \n \n---\n \n## Approved Bots\n \n---\n \n- [@L4s@lemmy.world](https://lemmy.world/u/L4s) \n- [@autotldr@lemmings.world](https://lemmings.world/u/autotldr) \n- [@PipedLinkBot@feddit.rocks](https://feddit.rocks/u/PipedLinkBot) \n- [@wikibot@lemmy.world](https://lemmy.world/u/wikibot) ","removed":false,"published":"2023-06-11T02:16:17.173483Z","updated":"2025-03-21T18:06:07.536094Z","deleted":false,"nsfw":false,"actor_id":"https://lemmy.world/c/technology","local":false,"icon":"https://lemmy.world/pictrs/image/0da8d285-3457-4e5b-af21-b38609b07eea.webp","banner":"https://lemmy.world/pictrs/image/be818d42-f96f-4a33-b963-2a6e35938d54.png","hidden":false,"posting_restricted_to_mods":false,"instance_id":24,"visibility":"Public"},"creator_banned_from_community":false,"banned_from_community":false,"creator_is_moderator":false,"creator_is_admin":false,"counts":{"post_id":23802822,"comments":63,"score":-8,"upvotes":37,"downvotes":45,"published":"2025-06-10T06:37:24.931523Z","newest_comment_time":"2025-06-10T18:52:59.734711Z"},"subscribed":"NotSubscribed","saved":false,"read":false,"hidden":false,"creator_blocked":false,"unread_comments":63}],"moderates":[]}