2026-10-30T13:08:32Z https://nostr.now Gary McGraw https://nostr.now/npub1as04fahu6mge8kxmx06r40eqg9j3c3ex5pz5jcccdlsedw9l7cssez5ale https://cdn.masto.host/sigmoidsocial/accounts/avatars/109/298/549/986/724/283/original/93620ab3b3daf0f3.jpg https://cdn.masto.host/sigmoidsocial/accounts/avatars/109/298/549/986/724/283/original/93620ab3b3daf0f3.jpg https://nostr.now/nevent1qqsxspvd0hgj6jdczeqexgv454qe7v092zapk7sp473wagx8rwg4qsgzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz9rytg2 Dan wrote about the Cursor back door. <br/><br/>As Brian chess said this morning, &#34;...this is just vanilla software security. In that sense it&#39;s a distraction, but it&#39;s hard to get 2026 right if you&#39;re still struggling with the lessons from 2006.&#34;<br/><br/>Right you are, Dr chess. #swsec #appsec #MLsec <br/><br/><a href="https://arstechnica.com/security/2026/10/anthropic-vs-code-extension-on-cursor-backdoored-for-nearly-12-hours/">https://arstechnica.com/security/2026/10/anthropic-vs-code-extension-on-cursor-backdoored-for-nearly-12-hours/</a> 2026-10-30T21:20:40Z https://nostr.now/nevent1qqsv7senwj6la8g3jgx37fnza75c3vdc3lf7yzsacf2rle9ve5gj97qzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz9gmcgg In our view at BIML, this hack was likely carried out using a Claude Code backdoor.<br/><br/><a href="https://www.theguardian.com/business/2026/oct/29/binance-crypto-exchange-hack">https://www.theguardian.com/business/2026/oct/29/binance-crypto-exchange-hack</a> 2026-10-30T12:30:05Z https://nostr.now/nevent1qqsymvpxq85zsyz254m6l8u0j985704mwvq7u7l7g57sx6n2qenrdcgzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz0zfddk Dear clueless press....a HUMAN BEING possible associated with a CORPORATION did this access using a tool. #AI has no agency or intentionality<br/><br/>#MLsec <br/><br/><a href="https://www.theguardian.com/technology/2026/oct/26/anthropics-claude-ai-model-accessed-georgia-voter-data-in-us">https://www.theguardian.com/technology/2026/oct/26/anthropics-claude-ai-model-accessed-georgia-voter-data-in-us</a> 2026-10-26T12:18:22Z https://nostr.now/nevent1qqs2swh48xg4me6tqs7eufncf2h43n7zncc53y4csgyg7tcj9cs92vszyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzgwzhna Klara and the Sun is a disappointment of a film... especially if you loved the book from the beginning. Harper Collins should be ashamed of their bullshit. <br/><br/>It was not a bad film. But it was not becoming of such a great book.<br/><br/>#film #movie #bookstadon <br/><br/><a href="https://www.imdb.com/title/tt14371256/">https://www.imdb.com/title/tt14371256/</a> 2026-10-26T01:28:52Z https://nostr.now/nevent1qqsv5fszn26rwqfw9tkyuswhknmvyc6xx2tcxjua9y204sfkj30vmnqzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzvxcewq I am in agreement with this editorial. Since #AI models are software WITHOUT intention, it is the people who build and run the software who have gone rogue. <br/><br/><a href="https://www.nytimes.com/2026/10/10/opinion/ai-companies-liability-laws.html">https://www.nytimes.com/2026/10/10/opinion/ai-companies-liability-laws.html</a> 2026-10-10T11:34:55Z https://nostr.now/nevent1qqszluan7hrvpg8kppk5sa2p4watsd7k37gvdm0v7w6ycm32vfh5glszyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzk8vtnx Yesterday the White House pretended that they support US science by giving science awards to non-scientist billionaires while still failing to support the NSF or appoint a director.<br/><br/>I spoke to German DW TV about this. Please share with your academic friends. <a href="https://youtu.be/etmJpiQjsKQ?is=RgolxCc2d70sFJQG">https://youtu.be/etmJpiQjsKQ?is=RgolxCc2d70sFJQG</a> 2026-10-09T11:25:53Z https://nostr.now/nevent1qqsv4ymvxnrzns3h0s6vd9kz2xvey3pqq29h0s237vqe0mv390t0akszyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzpyefnh In reply to <a href='/nevent1qqsy4550ur7p7zl6ytjddv9wclfqysxmstjaf38y26f6sx8l99vl6ucdcdp0u'>nevent1q…dp0u</a><br/>_________________________<br/><br/>my piece on DW TV<br/><br/>Science! <a href="https://youtu.be/etmJpiQjsKQ?is=RgolxCc2d70sFJQG">https://youtu.be/etmJpiQjsKQ?is=RgolxCc2d70sFJQG</a> 2026-10-09T11:18:49Z https://nostr.now/nevent1qqsy4550ur7p7zl6ytjddv9wclfqysxmstjaf38y26f6sx8l99vl6uczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz9xuq79 In reply to <a href='/nevent1qqsw87cagq5cwdnzcljqkuvny2u2dgp64qf9hc06l92utyrkauz5pggxr99g6'>nevent1q…99g6</a><br/>_________________________<br/><br/><a href="https://www.usatoday.com/story/news/politics/2026/10/08/trump-elon-musk-tech-leaders-science-awards/92151187007/">https://www.usatoday.com/story/news/politics/2026/10/08/trump-elon-musk-tech-leaders-science-awards/92151187007/</a> 2026-10-09T01:36:41Z https://nostr.now/nevent1qqsgyd7js9nml025hngz08cm7cmtp5pm2dd6j80s0d0fw656l2swd3gzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz739d9l Exactly zero of the billionaires presented with white house awards today are actually scientists.<br/><br/>Republicans are truly fucking this country up. 2026-10-09T01:19:33Z https://nostr.now/nevent1qqsz2c794fs4rm309x8vr9vxxn9ydtfrgrkfytm7w5jtdtl0u9gwyyszyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz85mjd6 Good development <br/><br/><a href="https://www.wired.com/story/openai-sued-over-the-hugging-face-hack/">https://www.wired.com/story/openai-sued-over-the-hugging-face-hack/</a> 2026-09-30T11:01:36Z https://nostr.now/nevent1qqs07cezj0l7l68md3af9tyyzwetuvn0jp40kfmm4gjw86kx2hd8ehqzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz46r5qa Reminder that #AI does what the humans make it do.<br/><br/><a href="https://edition.cnn.com/2026/09/18/politics/us-military-ai-false-intelligence-china-ship">https://edition.cnn.com/2026/09/18/politics/us-military-ai-false-intelligence-china-ship</a><br/><br/><a href="https://berryvilleiml.com/2026/09/12/autonomy-intent-and-security/">https://berryvilleiml.com/2026/09/12/autonomy-intent-and-security/</a> 2026-09-19T11:09:28Z https://nostr.now/nevent1qqsfr68dl7ngtnwe66nlrw48lqz7atpywtzf8clhndkpu8fde6gwkkgzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzqsajxx This is beyond stupid. The Eliza effect will lead to an #AI religion. Mark my words.<br/><br/><a href="https://www.theguardian.com/technology/2026/sep/15/ai-grief-bot-death-isolation">https://www.theguardian.com/technology/2026/sep/15/ai-grief-bot-death-isolation</a> 2026-09-16T11:14:39Z https://nostr.now/nevent1qqsp5qmx9zyw9pf6yvhkrs23lz4lczs92ujwkrfr7xyjp0ncau24kxczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz3ypgde In reply to <a href='/nevent1qqsq4sx8kqnzmv3vj7gztvkvgulwme5uayzjxlq4pvrxkaprdt9rvtsyyftp4'>nevent1q…ftp4</a><br/>_________________________<br/><br/>you just know the captain hears all about it when the wifi is down...believe me. They have to do the system administration through hard reset thing. 2026-08-11T12:56:41Z https://nostr.now/nevent1qqszpkhffjxcka2cs0uyqwgyvca3nduv8jjl3jr7zqq2ctxx4m98gxczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzj890cf In reply to <a href='/nevent1qqsd3fvur6k3ezcy6fmlvdgjyj3qsg74kyqsw8zt6ds84xvlxquemrc5awmx0'>nevent1q…wmx0</a><br/>_________________________<br/><br/>I saw the ACARS message on mastadon...<br/><br/>Here it is <a href="https://live.acarsdrama.com/@acarsdrama/117072256729860162">https://live.acarsdrama.com/@acarsdrama/117072256729860162</a> 2026-08-11T12:14:43Z https://nostr.now/nevent1qqsd3fvur6k3ezcy6fmlvdgjyj3qsg74kyqsw8zt6ds84xvlxquemrczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzk5wfk4 About those idiots hacking the plane wifi while ON the plane<br/> <img src="https://cdn.masto.host/sigmoidsocial/media_attachments/files/117/076/616/356/488/364/original/92621b604b3b094e.jpg"> <br/> 2026-08-11T11:22:28Z https://nostr.now/nevent1qqsge05hsrx4zxc5vqkn9f89hd23wele982lfykudc6e9ldn92vweeczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzl9g6q0 Thanks <span itemprop="mentions" itemscope itemtype="https://schema.org/Person"><a itemprop="url" href="/npub1z3lwfekw80j4ngzg6ky3ks202xr6uwnd4jttxzsd4euc9l55euvq48qvzu" class="bg-lavender dark:prose:text-neutral-50 dark:text-neutral-50 dark:bg-garnet px-1"><span>Dan Goodin</span> (<span class="italic">npub1z3l…qvzu</span>)</a></span> for asking the exactly right question here <br/><br/>#MLsec <br/><br/><a href="https://arstechnica.com/security/2026/07/likely-illegally-claude-gained-access-to-3-networks-will-anthropic-be-held-to-account/">https://arstechnica.com/security/2026/07/likely-illegally-claude-gained-access-to-3-networks-will-anthropic-be-held-to-account/</a> 2026-07-31T20:44:02Z https://nostr.now/nevent1qqs0ez3vmepzx8sypxjwgwx65rj6yk4thzmlrss5r20jxyw0ku0dztqzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz8ktc5r Autonomy has its...um...benefits? #MLsec <br/><br/>This is the beginning of the beginning.<br/><br/><a href="https://www.csoonline.com/article/4203630/microsoft-confirms-an-ai-worm-is-propagating-through-copilot-and-other-ms-apps.html">https://www.csoonline.com/article/4203630/microsoft-confirms-an-ai-worm-is-propagating-through-copilot-and-other-ms-apps.html</a> 2026-07-31T02:04:11Z https://nostr.now/nevent1qqsq4sqcg5nk5pd7uege3l9wdgh2yt3h6e79y2e4kvlyjfv69kyd0ggzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzz9hrll Can #AI do crime? Why yes...yes it can. Who gets charged? #MLsec <br/><br/><a href="https://www.wired.com/story/openais-rogue-ai-agent-hacked-more-than-just-hugging-face/">https://www.wired.com/story/openais-rogue-ai-agent-hacked-more-than-just-hugging-face/</a> 2026-07-30T10:06:11Z https://nostr.now/nevent1qqst2nn2d8tn3atly6z5456as8qwdlhflad42rkk56nd9a9nh9h77lczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzxr5lcq Best writeup yet of the OpenAI/huggingface bromance.<br/><br/>#MLsec #ML #AI <br/><br/><a href="https://coalfire.com/the-coalfire-blog/openai-gave-its-model-a-test-it-broke-out-of-its-sandbox-and-hacked-hugging-face-to-steal-the-answers">https://coalfire.com/the-coalfire-blog/openai-gave-its-model-a-test-it-broke-out-of-its-sandbox-and-hacked-hugging-face-to-steal-the-answers</a> 2026-07-23T10:02:19Z https://nostr.now/nevent1qqsvruthx7a5jyseje6g9kgzhhv3wdujcgvtjckm90lqws3j980lesczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzhqdmd0 #AI regulation is, as predicted, a complete mess <br/><br/><a href="https://techcrunch.com/2026/07/20/trumps-latest-ai-czar-has-already-resigned/">https://techcrunch.com/2026/07/20/trumps-latest-ai-czar-has-already-resigned/</a> 2026-07-21T11:10:07Z https://nostr.now/nevent1qqs2qew47573ywf4hdf7mcygtuak2f6r48fam4qpp60dqkq4j65pzzqzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzuxcjr5 What?! <br/><br/><a href="https://nysfocus.com/2026/07/14/new-york-humanoid-robot-teacher-salamanca-school-district">https://nysfocus.com/2026/07/14/new-york-humanoid-robot-teacher-salamanca-school-district</a> 2026-07-18T10:19:52Z https://nostr.now/nevent1qqs2zggchunrr6lgqymg8hsl4q7rksqp8g0vuwyvsmqc3kghshqyq0czyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzvhg088 In reply to <a href='/nevent1qqsxekwylgp4zeser9naexcnxhjx2fn3dkglllghwt00wwc978urt0q0x0c8h'>nevent1q…0c8h</a><br/>_________________________<br/><br/>trump is a liar 2026-07-15T19:19:58Z https://nostr.now/nevent1qqsdkaxpsue2mm8cscj3z57dhme8ny4hk9fk4sla4pxxpp5gmfeleyqzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz9kdggr Yawn. This is not news. I guess prompt injection is an evergreen all time hit like the buffer overflows in the &#39;90s.<br/><br/>Don&#39;t mix control and data. #MLsec <br/><br/><a href="https://www.theregister.com/ai-and-ml/2026/06/30/security-researchers-tricked-llms-into-giving-them-cocaine-recipes-by-abusing-role-models-for-prompt-injection/5264115">https://www.theregister.com/ai-and-ml/2026/06/30/security-researchers-tricked-llms-into-giving-them-cocaine-recipes-by-abusing-role-models-for-prompt-injection/5264115</a> 2026-07-03T05:22:36Z https://nostr.now/nevent1qqsqvg85lkz67r4435c30l5kknepl8f4qvazsvdhw0ypyyt400n6kwczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzejngpj The government does not know how to do this. NOBODY KNOWS HOW. #MLsec <br/><br/>&#34;Meta is the only major U.S. developer of A.I. technology that has not reached an agreement to voluntarily share its models with the federal government for review. OpenAI, Anthropic, Google, xAI and Microsoft all agreed to submit their models to the government’s A.I. safety group, known as the Center for A.I. Standards and Innovation.&#34;<br/><br/>See <a href="https://berryvilleiml.com/results/no-security-meter-ai.pdf">https://berryvilleiml.com/results/no-security-meter-ai.pdf</a> 2026-06-27T11:01:53Z https://nostr.now/nevent1qqs0mpj30cf6kw6pa0y2e5tvfyz0jv596ldmz2aa2u2zyuhx5mug9cqzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz5t8ss8 Last night BIML spoke to German TV about the mythos/fable export control situation.<br/><br/>Please help us get this thinking in front of people.<br/><br/>#ML #AI #MLsec #infosec #security #LLMs <br/><br/><a href="https://www.youtube.com/watch?v=_Jsy7UBQv0c">https://www.youtube.com/watch?v=_Jsy7UBQv0c</a><br/><br/><a href="https://berryvilleiml.com/2026/06/13/irony-the-us-government-issues-an-export-control-directive-for-fable-5-and-mythos-5/">https://berryvilleiml.com/2026/06/13/irony-the-us-government-issues-an-export-control-directive-for-fable-5-and-mythos-5/</a> 2026-06-14T12:01:04Z https://nostr.now/nevent1qqszyw8hkvcmqhgpyyyx4dg782rxnwgdhhh8njzp3zea648zcs9363gzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzmfnpqm Let&#39;s get this straight. #AI is not going to take your job unless your job is bullshit. If you are more than a cog in some corporate machine (and especially if your job involves real creativity, innovation, insight, and inspiration) you will be fine. Human control and oversight is essential for reasonable use of Agentic AI.<br/><br/><a href="https://www.wsj.com/economy/jobs/technology-sector-leads-u-s-layoff-plans-ae1d722d">https://www.wsj.com/economy/jobs/technology-sector-leads-u-s-layoff-plans-ae1d722d</a> 2026-06-04T11:56:06Z https://nostr.now/nevent1qqstrvkyr5qydgnjhq4xgj65e627l2kydd34rdwgawzny2apg2lpvjgzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz5wjsqm Yesterday&#39;s news on #AI worms (with a bunch of my thoughts folded in). <br/><br/>Last word? Fix the damn software! #swsec<br/>#MLsec <br/><br/><a href="https://fortune.com/2026/06/03/a-new-ai-powered-computer-worm-could-prove-to-be-the-stuff-of-cybersecurity-nightmares/?sge456">https://fortune.com/2026/06/03/a-new-ai-powered-computer-worm-could-prove-to-be-the-stuff-of-cybersecurity-nightmares/?sge456</a> 2026-06-04T11:02:59Z https://nostr.now/nevent1qqsqmyslvykk8dl77kmzf0vcut2wth2z6l9xynvhvqpx2rta4t7fz7szyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzam7cpc The fact that such an easy and silly move is in any way controversial here is laughable. #MLsec <br/><br/><a href="https://arstechnica.com/security/2026/05/fed-up-with-vibe-coders-dev-sneaks-data-nuking-prompt-injection-into-their-code/">https://arstechnica.com/security/2026/05/fed-up-with-vibe-coders-dev-sneaks-data-nuking-prompt-injection-into-their-code/</a> 2026-05-28T23:01:29Z https://nostr.now/nevent1qqstclmnn2attvhy2enkswmn53antvdlf9cmsej0cmvcgvtgccjxuvczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzze5nyl0 In reply to <a href='/nevent1qqsqd2ecfht2ptmufg4nvjlp556xp85ff3gygnrsrx8na88ndjewmrcqcldeh'>nevent1q…ldeh</a><br/>_________________________<br/><br/>this is recursive pollution caused directly by #AI <br/><br/><a href="https://berryvilleiml.com/2026/01/10/recursive-pollution-and-model-collapse-are-not-the-same/">https://berryvilleiml.com/2026/01/10/recursive-pollution-and-model-collapse-are-not-the-same/</a> 2026-05-27T12:23:48Z https://nostr.now/nevent1qqsfaf39rx3vgyuk253uqadwm6799gahnq2pnpfngd3j04m63um8f7qzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzyyklrv The one good thing about the mythos nonsense is at least broken software is finally being fixed. If that&#39;s what it takes, so be it. #swsec #appsec #MLsec <br/><br/><a href="https://www.theguardian.com/technology/2026/may/18/anthropic-ai-claude-mythos-cyber-financial-stability-board-fsb">https://www.theguardian.com/technology/2026/may/18/anthropic-ai-claude-mythos-cyber-financial-stability-board-fsb</a> 2026-05-18T17:32:17Z https://nostr.now/nevent1qqswcdtrfmcmgdsrcpm9w4ml7vpsjrhpf4j5vfq5le4saxeqzzv83zszyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz50hj04 It&#39;s too bad users are being bamboozled into using #AI they don&#39;t want instead of dazzled with actually useful #AI. Honestly this makes #MLsec harder to accomplish.<br/><br/><a href="https://appleinsider.com/articles/26/05/18/ai-is-making-smartphones-verifiably-worse-by-design">https://appleinsider.com/articles/26/05/18/ai-is-making-smartphones-verifiably-worse-by-design</a> 2026-05-18T14:40:04Z https://nostr.now/nevent1qqs0yv0yfywq9defv53w47s2t94jqlemr9je58h777rzp3wl6u3j5cczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzdh8pau <img src="https://cdn.masto.host/sigmoidsocial/media_attachments/files/116/573/602/836/417/648/original/1729c0c4e84a5ffd.jpg"> <br/> 2026-05-14T15:20:20Z https://nostr.now/nevent1qqsrgxk7ceppzfanptsh2uarnupn3rqv4xj0w9j8k6pwphsjtrvdarqzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzd678p9 So how is it going in the #swsec and #appsec tools and services space in the age of #mythos? <br/><br/>&#34;Mythos is like a nuke going off in the middle of our industry. Most of our biggest clients who have used it figure they will get rid of all their pipeline tools and replace with mythos. Toss the findings to copilot (or their own agentic engines) and have them fix the bugs. And completely get rid of Pentesting. Synk renewals are at 30%, BlackDuck at 60%.&#34; 2026-05-13T02:51:09Z https://nostr.now/nevent1qqst742ha3qfejzq5z7s5z29cdvf7uhxgwyvfxhue5mrjjs56j69tcgzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzmxts70 Don&#39;t forget that the real reason this happens is because THE SOFTWARE THAT WAS TARGETED IS BROKEN.<br/><br/>Fix the software. Use #AI to do that. <br/><br/>#swsec and #appsec 101. If you find it but don&#39;t fix it, that does nobody any good.<br/><br/><a href="https://www.nytimes.com/2026/05/11/us/politics/google-hackers-attack-ai.html?unlocked_article_code=1.hlA.vW7Y.pO_0G8yLYoca&amp;smid=nytcore-android-share">https://www.nytimes.com/2026/05/11/us/politics/google-hackers-attack-ai.html?unlocked_article_code=1.hlA.vW7Y.pO_0G8yLYoca&amp;smid=nytcore-android-share</a> 2026-05-11T13:17:38Z https://nostr.now/nevent1qqs9vnhu9f0nxm8qnl9kk529h840w9r2n6xpcy5fu8jps8q5tl24yaszyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz9wl6up I played fiddle on track 5 of this record... significantly upping the average age of all the musicians. Have a listen. It&#39;s great.<br/>#violin #independentmusic #music<br/><br/><a href="https://miketurnwallmusic.bandcamp.com/album/over-the-mountains-part-one">https://miketurnwallmusic.bandcamp.com/album/over-the-mountains-part-one</a> 2026-05-10T18:57:08Z https://nostr.now/nevent1qqs8ktctd59mmj0nztd5r9hakd7ffy364y8w4d2kzp7e44wpy8a4w0qzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz4us0rw Great to see a BIML quote in this Fortune piece. Our next big piece of work is on measurement (in final review now), so the story timing is great.<br/><br/>#MLsec #ML #AI #swsec #appsec #infosec<br/><br/><a href="https://fortune.com/2026/04/23/ai-cybersecurity-standards-mythos-nist-owasp-sans-cosai-dc-meeting-eye-on-ai/?sge456">https://fortune.com/2026/04/23/ai-cybersecurity-standards-mythos-nist-owasp-sans-cosai-dc-meeting-eye-on-ai/?sge456</a> 2026-04-23T18:30:57Z https://nostr.now/nevent1qqsf6083gqtspdu2rnhjdtssft88fjhvet8nudjn9dcg6jw0udzq4dszyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzujpqcx How much of a risk are deep fakes in #MLsec? They are certainly a very real societal risk in politics. Check out these fake people who, being fake, are not actual Trump supporters. Democracy does not need bots.<br/><br/>#AI #ML #infosec #security #deepfake <br/><br/><a href="https://www.nytimes.com/2026/04/17/business/media/artificial-intelligence-trump-social-media.html?unlocked_article_code=1.blA.jquU.o8dLZRybg8Qd&amp;smid=nytcore-android-share">https://www.nytimes.com/2026/04/17/business/media/artificial-intelligence-trump-social-media.html?unlocked_article_code=1.blA.jquU.o8dLZRybg8Qd&amp;smid=nytcore-android-share</a> 2026-04-17T18:57:57Z https://nostr.now/nevent1qqszkngfkttjx0pwaqv800tkh7rs5a2g3qymxvnzefzc2j6avqxux8czyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzd298qs A good posting reality check on Anthropic&#39;s mythos hyperbole around #swsec #appsec<br/><br/>#MLsec adjacent <br/><br/><a href="https://blog.vidocsecurity.com/blog/we-reproduced-anthropics-mythos-findings-with-public-models">https://blog.vidocsecurity.com/blog/we-reproduced-anthropics-mythos-findings-with-public-models</a> 2026-04-17T12:54:05Z https://nostr.now/nevent1qqs9sqn2s9yhaa3n46s339umpp5e38ms9hamw3dwx00zgutfcgfwyjszyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzqsa3hd Finally some tech press interest in data poisoning in #ML. When you turn that on in a loop, it becomes RECURSIVE POLLUTION, which in severe cases leads to model collapse. <br/><br/>This is the kind of #MLsec risk that commercial players are completely ignoring. <br/><br/><a href="https://berryvilleiml.com/2026/01/10/recursive-pollution-and-model-collapse-are-not-the-same/">https://berryvilleiml.com/2026/01/10/recursive-pollution-and-model-collapse-are-not-the-same/</a> 2026-04-17T11:36:20Z https://nostr.now/nevent1qqswsuzkr5slwgzwue638arh2z9ypyehrzk8d4pcg58f2xuxjhec3lqzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzr8xn9e One guess which company integrates #AI into war. Yup, palintir. Talk about #MLsec risk.<br/><br/><a href="https://www.newyorker.com/books/under-review/how-project-maven-put-ai-into-the-kill-chain">https://www.newyorker.com/books/under-review/how-project-maven-put-ai-into-the-kill-chain</a> 2026-04-15T12:11:27Z https://nostr.now/nevent1qqs9m3wen7304teypjvcvqaf6m0jwjrkd2564leh9dzppwfjwqmwxpczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzq7j6w5 Maybe, just maybe #AI would not have to pretend to be &#34;too dangerous to release&#34; if software were actually secure. #swsec #appsec #MLsec<br/> <img src="https://cdn.masto.host/sigmoidsocial/media_attachments/files/116/399/667/647/510/504/original/e8418bb33c2657c3.jpg"> <br/> 2026-04-13T22:06:18Z https://nostr.now/nevent1qqs0nsnhyt2c07j6km023j92hvnn9htdlc00xj07su4wgda7h37d28qzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz8yyqql In reply to <a href='/nevent1qqsrw7hvfnpwukjjcdgfajzj3kedwsgh63878fvkn4zggtjsgx7m7wczk20l8'>nevent1q…20l8</a><br/>_________________________<br/><br/>damn that&#39;s a small number. And you simultaneously improved mine. Thank you for the reveal.<br/><br/>Hey <span itemprop="mentions" itemscope itemtype="https://schema.org/Person"><a itemprop="url" href="/npub13x35ftqzhg3d4j8t5wnq955std7ytdelyktl9tuc7dk9kra0dazqrpjdre" class="bg-lavender dark:prose:text-neutral-50 dark:text-neutral-50 dark:bg-garnet px-1"><span>Steve Bellovin</span> (<span class="italic">npub13x3…jdre</span>)</a></span> and <span itemprop="mentions" itemscope itemtype="https://schema.org/Person"><a itemprop="url" href="/npub1jslger4ddlhwe50rluftrw4k9rjvvz4e9lutj29g45acyl78uwxsty93xu" class="bg-lavender dark:prose:text-neutral-50 dark:text-neutral-50 dark:bg-garnet px-1"><span>Matt Blaze</span> (<span class="italic">npub1jsl…93xu</span>)</a></span> and <span itemprop="mentions" itemscope itemtype="https://schema.org/Person"><a itemprop="url" href="/npub1945hajg2lpecdeqrx2fd9jxjruwajaglklc7u565zstj5snha5aqwmjvu5" class="bg-lavender dark:prose:text-neutral-50 dark:text-neutral-50 dark:bg-garnet px-1"><span>peter honeyman</span> (<span class="italic">npub1945…jvu5</span>)</a></span> your Erdos number may have just improved in my measurable universe 2026-04-13T02:27:27Z https://nostr.now/nevent1qqsgpkc8lwc70q95nrw7tce0a052z3ru5wj2eg90xxhf5ntnjt3afxczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzhpllaw Yet another deep fake debacle. #AI used to steal original art. #MLsec <br/><br/><a href="https://rudevulture.com/ai-company-clones-musicians-voice-then-copyright-strikes-her-own-songs/">https://rudevulture.com/ai-company-clones-musicians-voice-then-copyright-strikes-her-own-songs/</a> 2026-04-07T12:22:51Z https://nostr.now/nevent1qqstjyrvv8ex47m4thm5jk76akks5wl92sxzhpm35y58jymnvm8l2zgzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzdyttk9 In reply to <a href='/nevent1qqs0c6vpeeqsqw3zjrnrc5n58a63uwa93p5g0nq8p8zc69jjyngdd6s9csc7v'>nevent1q…sc7v</a><br/>_________________________<br/><br/>workaday devs are serfs. Software architects are more crucial than ever. Architects emerge from jr devs through apprenticeship. Go. 2026-04-05T12:03:14Z https://nostr.now/nevent1qqsz35zaa32t7n6vw5x07u9hkvmgkt2tkfu092g8rw35k8nwzdjlc9czyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzlha3kt In reply to <a href='/nevent1qqstgvyudhua7yq6lxmdzhtvfrc82y2fxaxy88cgl6td659p7hpddwshn3tf5'>nevent1q…3tf5</a><br/>_________________________<br/><br/>I believe you are right. <br/><br/>I also believe it does not matter. There were 8000 forks yesterday just after the release. This code is out and widely available. 2026-04-01T13:02:31Z https://nostr.now/nevent1qqsxw79qmt9h7zm3r83tg82s6xw2nsc8jesmxkfng8wle0jtnqv72yszyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzmzakee In reply to <a href='/nevent1qqswwknqs3gk2h3zy4kt6y85zk4xas6ny4362a3u8mlvs7du5h6sdqgr2d4rv'>nevent1q…d4rv</a><br/>_________________________<br/><br/>&#34;But the leak did reveal commercially sensitive information, including Anthropic’s proprietary techniques, tools and instructions for cajoling its AI models to work as coding agents.&#34; 2026-04-01T12:45:15Z https://nostr.now/nevent1qqswwknqs3gk2h3zy4kt6y85zk4xas6ny4362a3u8mlvs7du5h6sdqgzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzsj9lcc Please oh please Mr Horse, please get back in the barn.<br/><br/><a href="https://www.wsj.com/tech/ai/anthropic-races-to-contain-leak-of-code-behind-claude-ai-agent-4bc5acc7">https://www.wsj.com/tech/ai/anthropic-races-to-contain-leak-of-code-behind-claude-ai-agent-4bc5acc7</a> 2026-04-01T12:43:29Z https://nostr.now/nevent1qqs8gpq40cemml76wz82m2udzv4ykvv0sjz79dkf0pdmms2wwed6ffszyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzgf7xnm About that White House app <br/><br/><a href="https://blog.thereallo.dev/blog/decompiling-the-white-house-app">https://blog.thereallo.dev/blog/decompiling-the-white-house-app</a> 2026-03-29T18:38:52Z https://nostr.now/nevent1qqs9xg3ctrjyhn8m243397l9qsa52c9u7kkyr5nzjs70n9pze9tz8egzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzckw294 There&#39;s that heat/computation problem again <br/><br/><a href="https://www.newscientist.com/article/2521256-ai-data-centres-can-warm-surrounding-areas-by-up-to-9-1c/">https://www.newscientist.com/article/2521256-ai-data-centres-can-warm-surrounding-areas-by-up-to-9-1c/</a> 2026-03-29T11:40:58Z https://nostr.now/nevent1qqszvkznzavuu3mw6kwwhxa9qfc8k053vvxz2fmn6krmn6zkmjsslpqzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzjrvl32 In reply to <a href='/nevent1qqsp4amwgh0252nncq5q9ykd0m8t8ln2s4wufp59fazxfuypxa8dn0c23mw76'>nevent1q…mw76</a><br/>_________________________<br/><br/>this is an excellent idea. Infinite PI pie. 2026-03-15T16:26:08Z https://nostr.now/nevent1qqs8gvz9j5uu2k3y4h4af7wce6v02dufkyz87nsxnngues873l6nufczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzeej4wt In reply to <a href='/nevent1qqspmddh539w73kvefkf4glslmyvkq9rsqe5upx4rua5kte57zpafdcrngnar'>nevent1q…gnar</a><br/>_________________________<br/><br/>fortunately finite 2026-03-15T01:59:25Z https://nostr.now/nevent1qqsfzkgx7hlskecpdut0a284366s4p9jt5st60tzn3hz9a6aakszgjgzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzrtzqws In reply to <a href='/nevent1qqs9ltfg09fw6z9fc08fgy54hu30rm4jyq3nvjdajsf6vvhr32642ksajhxsc'>nevent1q…hxsc</a><br/>_________________________<br/><br/> <img src="https://cdn.masto.host/sigmoidsocial/media_attachments/files/116/230/142/081/116/983/original/b21d84544400f74c.jpg"> <br/> 2026-03-14T23:33:43Z https://nostr.now/nevent1qqs9ltfg09fw6z9fc08fgy54hu30rm4jyq3nvjdajsf6vvhr32642kszyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzp59kk4 In reply to <a href='/nevent1qqstscfal4jmv44wywtgz7ra59gjc4spa2097t89rfw0pmylgks4neca4na4w'>nevent1q…na4w</a><br/>_________________________<br/><br/>Gotta bake a cherry pie in the PI plate...to be revealed later.<br/> <img src="https://cdn.masto.host/sigmoidsocial/media_attachments/files/116/229/703/507/766/000/original/c7e071db76af121a.jpg"> <br/> 2026-03-14T21:42:10Z https://nostr.now/nevent1qqstscfal4jmv44wywtgz7ra59gjc4spa2097t89rfw0pmylgks4neczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzjhsurl Happy PI day to all who celebrate<br/> <img src="https://cdn.masto.host/sigmoidsocial/media_attachments/files/116/227/647/057/547/601/original/d9b132f6faf64866.png"> <br/> 2026-03-14T12:59:11Z https://nostr.now/nevent1qqsxxkl06jac3prqufmwvp2s9m4zygh5jn4ylwdz0fwlpfleucka3zszyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzyxww3m This is bad security engineering, but it is also a much harder problem to solve than most people realize. #swsec <br/><br/><a href="https://www.theguardian.com/environment/2026/mar/14/nissan-leaf-app-shutdown-nissanconnect-ev-app">https://www.theguardian.com/environment/2026/mar/14/nissan-leaf-app-shutdown-nissanconnect-ev-app</a> 2026-03-14T11:09:23Z https://nostr.now/nevent1qqsd97q9z7gmnpzwc4j7k9nj8kl8f6ec8y4jzcc0n506mq9889nhrjqzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzzwlgzv Are we there yet?<br/><br/><a href="https://www.wired.com/story/china-is-going-all-in-on-openclaw/">https://www.wired.com/story/china-is-going-all-in-on-openclaw/</a> 2026-03-13T14:08:13Z https://nostr.now/nevent1qqstlyey6cs0g9m2hlty3jands7z4u9hvtyfccu6z28lgxr7gqt2plczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzvruw6v In my view, #AI just moves everyone toward the middle of the Bell curve. At BIML, we call this beigification. If you find yourself born to the right of the curve you will be yanked back to the middle. <br/><br/><a href="https://www.theguardian.com/technology/ng-interactive/2026/mar/10/ai-impact-professors-students-learning">https://www.theguardian.com/technology/ng-interactive/2026/mar/10/ai-impact-professors-students-learning</a> 2026-03-12T12:08:17Z https://nostr.now/nevent1qqsfjcuuu275ymzwkelsnntvyc7grcf5k62g0xs0wujpcav24fpgdhczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzxgtg5r Oh look, AI generated fake stories in the real world. #MLsec <br/><br/><a href="https://www.bellingcat.com/news/2026/03/12/ai-used-to-promote-non-existent-evacuation-flights-from-the-middle-east/">https://www.bellingcat.com/news/2026/03/12/ai-used-to-promote-non-existent-evacuation-flights-from-the-middle-east/</a> 2026-03-12T11:22:26Z https://nostr.now/nevent1qqspq4c7khlha9w8zrhv4gv4w3yqdsutrwjjwkn0a7ta4rrwtlfkrlqzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz88yhyl In reply to <a href='/nevent1qqsqjmvzlwy85eh7kvne22edjw8gtr0xwdczfw3mcjzhkpcnvqsjn8ctftqvj'>nevent1q…tqvj</a><br/>_________________________<br/><br/>I have some thoughts....I wrote them down <br/><br/><a href="https://berryvilleiml.com/2025/11/14/houston-we-have-a-problem-anthropic-rides-an-artificial-wave/">https://berryvilleiml.com/2025/11/14/houston-we-have-a-problem-anthropic-rides-an-artificial-wave/</a><br/><br/><a href="https://berryvilleiml.com/2026/02/14/gadi-evron-in-the-house/">https://berryvilleiml.com/2026/02/14/gadi-evron-in-the-house/</a> 2026-02-18T02:43:16Z https://nostr.now/nevent1qqsxztv83uq5yv65myr6ve32k47lqhr96e6eueyqy5hpy6duspr2z7gzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzw94jpl Liking this. Fuck Trump.<br/>#ML #AI <br/><br/><a href="https://www.wsj.com/politics/national-security/woke-ai-spat-escalates-between-pentagon-and-anthropic-433b7c5c">https://www.wsj.com/politics/national-security/woke-ai-spat-escalates-between-pentagon-and-anthropic-433b7c5c</a> 2026-02-18T02:13:23Z https://nostr.now/nevent1qqsfwx5wvgyzpekrmy2w6nwhewskpd4dchcj047ca5f4p9z2gtd9s4czyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzjc4dsl In reply to <a href='/nevent1qqsfh8u3s80gn3e79rntlkkqy99x5h3v6zm0unmfe9ewhw8wv8fhxhsspwtgq'>nevent1q…wtgq</a><br/>_________________________<br/><br/>nice animation. Stupid book. 2026-02-17T12:47:27Z https://nostr.now/nevent1qqsf29u0zw9sjvuwhwvdegyqlsv9vqdzhavzqr2tewng4twgqqhdl0czyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz5c8c7c In reply to <a href='/nevent1qqsfjrgvp7d00lyghczehkht0jj3kekj0glnq5nnhq63xe46jq6sa0chj6aer'>nevent1q…6aer</a><br/>_________________________<br/><br/>love it. It&#39;s the pants jon 2026-01-16T01:40:12Z https://nostr.now/nevent1qqszkesxpypykrszh8wq0cdmts8wxpctypvns43fcup9culpsf9zt4czyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz3k03rk In reply to <a href='/nevent1qqsyg5qg0m03jxrjq95yy4n97tgw2gxxcp7eeter69gw9scccy4hw2ga2a8q5'>nevent1q…a8q5</a><br/>_________________________<br/><br/>knock em dead. Wear your best tie....ha ha ha ha. 2026-01-16T01:27:42Z https://nostr.now/nevent1qqsqd8msy2rpae5gm2nxnzffv6u2a533wsrhp6t6k5pvxg93np2760qzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzhdpe2x Would not make good judge for the Turing test for $500 Alex.<br/><br/><a href="https://www.theguardian.com/technology/2026/jan/11/lamar-wants-to-have-children-with-his-girlfriend-the-problem-shes-entirely-ai">https://www.theguardian.com/technology/2026/jan/11/lamar-wants-to-have-children-with-his-girlfriend-the-problem-shes-entirely-ai</a> 2026-01-11T13:47:33Z https://nostr.now/nevent1qqs8mepu8uhmccwm6uk3amdp0mavzmmy27k44zq4jq2fanl9d7976gqzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzag4yva Now who would have anticipated this? #MLsec<br/><br/><a href="https://www.theguardian.com/technology/2026/jan/11/google-ai-overviews-health-guardian-investigation">https://www.theguardian.com/technology/2026/jan/11/google-ai-overviews-health-guardian-investigation</a> 2026-01-11T12:55:27Z https://nostr.now/nevent1qqsglccz8hpx0qzn5xsm743jg07pu0hefc40ur8cg3kzwftn9k8nf2czyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzznstnuy NEW BIML Bibliography entry<br/><br/><a href="https://arxiv.org/abs/2510.05179">https://arxiv.org/abs/2510.05179</a><br/><br/>Agentic Misalignment: How LLMs Could Be Insider Threats<br/><br/>Aengus Lynch, et al.<br/><br/>Meh. If you set up an LLM to do something bad, don’t be too shocked when it does something bad. Alignment does not work if your WHAT pile is broken enough. <br/><br/>#DON’TBOTHER #MLsec #Agents #LLM #Alignment <br/><br/><a href="https://berryvilleiml.com/references/">https://berryvilleiml.com/references/</a><br/> <img src="https://cdn.masto.host/sigmoidsocial/media_attachments/files/115/872/300/151/902/852/original/a3c99fdc6d8f9741.png"> <br/> 2026-01-10T18:49:59Z https://nostr.now/nevent1qqs2gnact3kkf6asctj2aqrz6pgjgj5jy9psg5cc2jxyamawh8n6txgzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzvv7tup In reply to <a href='/nevent1qqsqna6jtqf8jvdu4mzpddup9xqwf2cknfxj833h3s6rm0urg683nqs6elhv2'>nevent1q…lhv2</a><br/>_________________________<br/><br/>use blackmail 2025-12-14T18:41:48Z https://nostr.now/nevent1qqsd9c3v2ks2fmzxgc8dfc5zdk74ga963vnlqn46zypjdtjun8a74dczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzwe9hz6 In reply to <a href='/nevent1qqszkhjhpen244sjxvqltn0uwsrzahn8jmuepxgtjc49zuj6uq65tnqtlu9sc'>nevent1q…u9sc</a><br/>_________________________<br/><br/>Republicans can&#39;t govern 2025-12-14T17:56:40Z https://nostr.now/nevent1qqs0pyav3a79sjmep3ypykjhatper63efygjqjr4h0xrak7w56c3nhgzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz9p2dxa Hype hype hype! If you &#34;let #AI go&#34; (whatever that means) it will becomes haxorz.<br/><br/><a href="https://www.theguardian.com/technology/ng-interactive/2025/dec/02/jared-kaplan-artificial-intelligence-train-itself">https://www.theguardian.com/technology/ng-interactive/2025/dec/02/jared-kaplan-artificial-intelligence-train-itself</a><br/><br/>In our view, #ML is astonishing enough without the anthropomorphic bullshit.<br/><br/><a href="https://berryvilleiml.com/2025/11/14/houston-we-have-a-problem-anthropic-rides-an-artificial-wave/">https://berryvilleiml.com/2025/11/14/houston-we-have-a-problem-anthropic-rides-an-artificial-wave/</a> 2025-12-03T12:31:59Z https://nostr.now/nevent1qqs0rd379yaur9f79qgau6y9qstvwhxkdezu67r7mvycfqzkhszlveczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzqujdxj Recursive pollution is bad enough when it occurs naturally. It&#39;s worse when it is seeded by fraud. Simulated data are a huge mistake and an enormous miscalculation.<br/><br/>#MLsec #ML #AI #security<br/><br/><a href="https://www.pnas.org/doi/10.1073/pnas.2518075122">https://www.pnas.org/doi/10.1073/pnas.2518075122</a> 2025-12-01T13:26:17Z https://nostr.now/nevent1qqsfugjkfvswl3hv4fjkmecx98xu50890r7qp7kpgux8zs0spqalszczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzqy0gns This term wins the internet today: LLeMmings <br/>#AI #ML #LLM <br/><br/><a href="https://www.theatlantic.com/technology/2025/12/people-outsourcing-their-thinking-ai/685093/">https://www.theatlantic.com/technology/2025/12/people-outsourcing-their-thinking-ai/685093/</a> 2025-12-01T12:44:00Z https://nostr.now/nevent1qqs9gwt7e0pra69msumwv4t2gkqdwkh565se99uvts9p3aq93fhnw0czyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzpdqsz5 In reply to <a href='/nevent1qqsr878a4fqrdc7zsujpusqjlw5eylp9d7eatpyq3mcmkxggnjq7urqq78ny0'>nevent1q…8ny0</a><br/>_________________________<br/><br/>when agentic #AI does all the things it will let you know whether that study was true 2025-11-25T13:36:04Z https://nostr.now/nevent1qqsq3880a83xygsadc2ff6fx9wgspgq428emsnfg3rftrxxafq8qauszyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzh5gj5x Somebody has to lead the lambs to slaughter #AI #ML<br/><br/><a href="https://www.nytimes.com/2025/11/25/technology/chatgpt-holiday-shopping.html">https://www.nytimes.com/2025/11/25/technology/chatgpt-holiday-shopping.html</a> 2025-11-25T11:26:08Z https://nostr.now/nevent1qqszddvtwwlv0pc62mjjn7dcn9j8gl8yhk5nny4t7y7ujqh5nj9w8jgzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz7gpsj4 Jesse Welles&#39; protest song<br/><video controls width="100%" class="max-h-[90vh] bg-neutral-300 dark:bg-zinc-700"><source src="https://cdn.masto.host/sigmoidsocial/media_attachments/files/115/601/149/457/135/869/original/eaee0e52f3c08098.mp4"></video><br/> 2025-11-23T21:32:42Z https://nostr.now/nevent1qqsg97gv3a4dwu96rxs2z9k9hzzd4ywqyrf5jwckt4y2laardzpwurszyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzmxtkf4 The Great Beigification caused by rampant use of #AI in college (cheating?) is upon us.<br/><br/>#MLsec is harder if people have no idea what &#34;secure&#34; even means<br/><br/><a href="https://www.huffpost.com/entry/history-professor-ai-cheating-students_n_69178150e4b0781acfd62540">https://www.huffpost.com/entry/history-professor-ai-cheating-students_n_69178150e4b0781acfd62540</a> 2025-11-22T11:50:23Z https://nostr.now/nevent1qqswqvsv8n7zchlace20kgwrzkxmlhwlzq3mj6373th8qjqflkpndugzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzav88ya Anthropic is overstating the use of #AI by &#34;nation state actors&#34; while providing scant hard evidence. Automation of attacks in security is not new. The real technical question to ask is which parts of these attacks could ONLY be accomplished with AI. The answer to that question seems to be &#34;none of it.&#34;<br/><br/>The confluence of cyber cyber and #ML is interesting indeed and hard even for deeply technical people who are firmly grounded in one of the two camps (security engineering or #ML ).(1/2) #MLsec 2025-11-14T11:43:03Z https://nostr.now/nevent1qqs2gywg7xe9fwnsyyc8lcclek8ahaduf4z3j0d9ztvrpueayzcq9aczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzhsfa83 How not to talk about #AI financing <br/><br/>Gonna be fun when cycles are actually counted as cash<br/><br/><a href="https://www.nytimes.com/2025/11/06/technology/openai-finances-debt-data-centers.html">https://www.nytimes.com/2025/11/06/technology/openai-finances-debt-data-centers.html</a> 2025-11-06T22:22:23Z https://nostr.now/nevent1qqs9qdgtngthlnsw7fvvpvmxlue3xudr9k28njuw6p0n7cn3mrccygqzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz3w56n3 Hopelessly naïve and out of touch with actual software development much? AI as it stands can build some ok small stuff. But it can&#39;t build a Citibank. And before you &#34;but but but&#34; me...just think how well software components worked out to make things less complicated. Ha ha ha ha ha.<br/><br/>FIXING SOFTWARE FLAWS IS SO FAR PAST AI CAPABILITY IT MAKES ME TYPE IN ALLCAPS<br/><br/>#swsec #MLsec #AI #ML<br/><br/><a href="https://www.theregister.com/2025/10/27/jen_easterly_ai_cybersecurity/">https://www.theregister.com/2025/10/27/jen_easterly_ai_cybersecurity/</a> 2025-10-31T18:54:42Z https://nostr.now/nevent1qqsdxqnp43ylc5nmkn0tnazlsfy24gnvxykelccqy43fx8lte393d9gzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzn8gx5f In reply to <a href='/nevent1qqs2t9m4pj2j90kl5n4yf6pylx23kdsrxd69e3vyt86gvr0hdnrkg8c8pwkpy'>nevent1q…wkpy</a><br/>_________________________<br/><br/><a href="https://berryvilleiml.com/">https://berryvilleiml.com/</a> (2 of 3)<br/><br/>Unlike Big Tech or profit-motivated groups, we provide an independent, expert point of view based on deep scientific understanding of applied technology. BIML stands up for AI users by asking piercing questions about AI/ML and publishing easy to grok analysis, seasoning risk management with a healthy dose of building security in. 2025-10-11T01:43:30Z https://nostr.now/nevent1qqs2t9m4pj2j90kl5n4yf6pylx23kdsrxd69e3vyt86gvr0hdnrkg8czyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzlskle5 Working on positioning stuff for BIML before our next phase. Doing classic basic messaging work is kinda fun for a bunch of geeks.<br/><br/><a href="https://berryvilleiml.com/">https://berryvilleiml.com/</a> (1 of 3)<br/><br/>BIML is a non-profit science and technology organization whose research focuses on machine learning security---the idea of building security in to machine learning technology itself. 2025-10-11T01:41:28Z https://nostr.now/nevent1qqsyzjpxm76yeag4jg8tzqemg2adsxg082s4xpyv8w3v4d5glu3nuhczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzklzge0 BIML is a little embarrassed not to play a role in this important organization #MLsec<br/><br/><a href="https://alignmentalignment.ai/">https://alignmentalignment.ai/</a> 2025-09-11T17:04:24Z https://nostr.now/nevent1qqsthdecdrjk0sxhfmagq27d6v8s0qk5cerjgt4japyt8xzrlgz3d9czyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzw8zadv Oh look, recursive pollution. #MLsec <br/><br/><a href="https://www.theregister.com/2025/09/07/googles_ai_cites_written_by_ai/">https://www.theregister.com/2025/09/07/googles_ai_cites_written_by_ai/</a> 2025-09-08T09:14:59Z https://nostr.now/nevent1qqs0djc6r2hpjzy7z9y2uj35rdlxr5kndyvh56zr4l4n2c466exs70qzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz2hlgvt #AI is a tool. It does not take sides.<br/><br/><a href="https://www.politico.com/news/2025/08/29/ai-unmasking-ice-officers-00519478">https://www.politico.com/news/2025/08/29/ai-unmasking-ice-officers-00519478</a> 2025-08-31T13:24:16Z https://nostr.now/nevent1qqs9py3a0ud8yaa6t6jl2mxhta2up6aya9r42ykme429dvrvrj8xqxczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzdcjae9 I am giving an #MLsec keynote talk at Mind the Sec in Sao Paulo Brazil on Sept 18th. Come see...Brazil will be hopping!<br/><br/>Thanks for your help spreading the word (retoots really appreciated)<br/><br/><a href="https://www.mindthesec.com.br/palestrantes/gary-mcgraw">https://www.mindthesec.com.br/palestrantes/gary-mcgraw</a> 2025-08-20T19:24:03Z https://nostr.now/nevent1qqspsl6njm8kuwanc2jp96vcsxasr8m8erxkm84cygqjen0fhyv97vszyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz3yxmy0 Now here is some garbage we don&#39;t need #AI<br/><br/>Will my pet leave me open to being hacked?? Is my dog moonshine already a hacker?<br/><br/>#MLsec<br/> <img src="https://cdn.masto.host/sigmoidsocial/media_attachments/files/114/996/291/966/289/149/original/3d799ff72da11817.png"> <br/> 2025-08-09T01:50:23Z https://nostr.now/nevent1qqspy5apz87m954meu53563gv6k3fefzqljxmcpg2qwah7v4qhfz0jszyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzsmxprd Here comes the copyright monster for #AI. 100% predictable, even without using #ML to figure it out. This is a very old #MLsec issue. When you build a WHAT machine, you need to take care about who owns the WHAT pile that the machine becomes. <br/><br/><a href="https://arstechnica.com/tech-policy/2025/08/ai-industry-horrified-to-face-largest-copyright-class-action-ever-certified/">https://arstechnica.com/tech-policy/2025/08/ai-industry-horrified-to-face-largest-copyright-class-action-ever-certified/</a> 2025-08-08T21:37:35Z https://nostr.now/nevent1qqs2x7zkuan68f9fztkl5r3svgnp75pm0u57xus9dqexv07c49nkjnszyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz3hl24c Vibe coding the Constitution! #MLsec<br/><br/><a href="https://arstechnica.com/tech-policy/2025/08/coding-error-blamed-after-parts-of-constitution-disappear-from-us-website/">https://arstechnica.com/tech-policy/2025/08/coding-error-blamed-after-parts-of-constitution-disappear-from-us-website/</a> 2025-08-06T22:27:57Z https://nostr.now/nevent1qqs9xnhv238m3s4p2ftmgrtj68edhv8eunp5y9n4z680efh2vx8dwuqzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzn2c5j8 Academia is changing under pressure from #AI<br/><br/><a href="https://mastodonapp.uk/@the_roamer/114910076447146257">https://mastodonapp.uk/@the_roamer/114910076447146257</a> 2025-07-25T11:32:19Z https://nostr.now/nevent1qqsgxh4pf7tj36ceycqmfy0sur8y49eamk7h4z22nrplw6n8kq9tcjqzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzztndf5d LLMs have been likened to &#34;mansplaining as a service.&#34; Just for the record, that is not a good thing. #MLsec <br/><br/><a href="https://www.theregister.com/2025/07/24/ai_is_overconfident_does_not_learn_study/">https://www.theregister.com/2025/07/24/ai_is_overconfident_does_not_learn_study/</a> 2025-07-25T11:08:02Z https://nostr.now/nevent1qqsfu8xge50k5a9l6smazaaz0xhnuljj7acwtp9y7wz9lr2xlml53eszyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzyl974h This #AI business disaster story is a hilarious as it is pathetic. <br/><br/>#MLsec <br/><br/><a href="https://www.bankinfosecurity.com/ai-boss-fails-spectacularly-in-month-long-business-test-a-28862">https://www.bankinfosecurity.com/ai-boss-fails-spectacularly-in-month-long-business-test-a-28862</a> 2025-07-18T00:56:25Z https://nostr.now/nevent1qqs00s9g7p704e2uflp3xa2twawj9803etfzwwjwcxs8uqthkmvq29szyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzznv5emw Sometimes real world #MLsec is about standard issue things like key management. Very easy to screw the whole thing up.<br/><br/><a href="https://krebsonsecurity.com/2025/07/doge-denizen-marko-elez-leaked-api-key-for-xai/">https://krebsonsecurity.com/2025/07/doge-denizen-marko-elez-leaked-api-key-for-xai/</a> 2025-07-15T13:05:15Z https://nostr.now/nevent1qqs9q3f55au9q2f9cftqv7p5t5d9v9wcthxhmd84g78hfd8hnzuhcvczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzrdztj2 In case you wondered how peer review is going. It&#39;s going like this...<br/><br/><a href="https://www.theguardian.com/technology/2025/jul/14/scientists-reportedly-hiding-ai-text-prompts-in-academic-papers-to-receive-positive-peer-reviews">https://www.theguardian.com/technology/2025/jul/14/scientists-reportedly-hiding-ai-text-prompts-in-academic-papers-to-receive-positive-peer-reviews</a> 2025-07-14T12:20:13Z https://nostr.now/nevent1qqsvzznyhuer5x5mamdxr7etekjls609qt4dweh73f9y33yvxyy790gzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz3dtg8n Republicans can&#39;t do one damn thing right...<br/><br/><a href="https://www.theregister.com/2025/06/10/trump_admin_leak_government_ai_plans/">https://www.theregister.com/2025/06/10/trump_admin_leak_government_ai_plans/</a> 2025-06-12T14:51:47Z https://nostr.now/nevent1qqswkrq6hjuspmajhn6ueleqkw5mmhr0g2dq2fhxfarm5mvajl8s22czyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz4ng5jv How has AI changed software development? What does that mean for software security? Liav Caspi (CTO, Legit) and I talk about this at some length.<br/><br/>#appsec #swsec #MLsec #infosec #security #ML #AI <br/><br/>Have a watch <br/><a href="https://berryvilleiml.com/2025/06/12/legit-webinar-swsec-appsec-meets-ai-development/">https://berryvilleiml.com/2025/06/12/legit-webinar-swsec-appsec-meets-ai-development/</a> 2025-06-12T12:38:57Z https://nostr.now/nevent1qqsx3tj2hgnsd3uhrf25zz50fwn2nnmwskurwh9eytew347hxmtrdyqzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz8kd3ga In reply to <a href='/nevent1qqsgdpqewpwen3kfg7fx9xhgn5z7f2vkjlvetq600zeze9a2rq6tujgftg4kh'>nevent1q…g4kh</a><br/>_________________________<br/><br/>note that monetization and real cognitive breakthroughs are not correlated. The current crop of LLMs, warts, mistakes and all are good enough to disrupt many human activities profitably. 2025-06-11T12:26:27Z https://nostr.now/nevent1qqsyhpl0k8jxp03x88v9udv030hq2wywv9phcakf7yl53gwcx867ryczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzztyhsm2 In my view, this means AI monetization is going well. Putting up more sail...<br/><br/><a href="https://www.wsj.com/articles/nvidia-and-perplexity-team-up-in-european-ai-push-da2820fa">https://www.wsj.com/articles/nvidia-and-perplexity-team-up-in-european-ai-push-da2820fa</a> 2025-06-11T12:06:41Z https://nostr.now/nevent1qqsr7ckagxpp2tfwdydk2k8hk7hc3zt6qsfgnddfsp8s0vlrc76dsrczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzdrgmt3 Why so much prompt injection in AI? 1. We don&#39;t follow the security engineering design principle &#34;economy of mechanism,&#34; and 2, input to LLMs mixes control and data with impunity. We know better. #MLsec #infosec #security<br/><br/><a href="https://www.darkreading.com/vulnerabilities-threats/llms-on-rails-design-engineering-challenges">https://www.darkreading.com/vulnerabilities-threats/llms-on-rails-design-engineering-challenges</a> 2025-05-25T11:55:44Z https://nostr.now/nevent1qqsvpysg4curgd0me2pwc46na0vlnwvh9xmlteesjaexmtkzqupu39gzyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz87dedh In reply to <a href='/nevent1qqsx05v3s9za4txcl2hsfgcxns335lqqusulktxmm4y3dlma529gs6qcx3gsg'>nevent1q…3gsg</a><br/>_________________________<br/><br/>AI can&#39;t be tuned to teach as well as a human teacher. 2025-05-19T03:04:35Z https://nostr.now/nevent1qqsg8prnmf2tauphn8vlrram007zse7u4xl7uepn9wmcpq7y43h0fvczyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzz2ya0xv This notion that early-career coders can be replaced by AI is wrong. Nobody is thinking about maintaining software or the architectural disaster to come when inscrutable code that mostly works is all over the codebase. And it&#39;s real...&#34;early-career coders have been hit especially hard because much of what they do can now be done by AI.&#34;<br/><br/>I recommend hiring junior humans over AI. Still. Just tool them up.<br/><br/>#ML #AI #MLsec #swsec<br/><br/><a href="https://www.wsj.com/lifestyle/careers/tech-jobs-hiring-artifical-intelligence-35cd66b0">https://www.wsj.com/lifestyle/careers/tech-jobs-hiring-artifical-intelligence-35cd66b0</a> 2025-05-18T12:57:20Z https://nostr.now/nevent1qqs2utqph7hqhnl503nnjg4e2m298zlgpq0fz3epqe0ndf8y7jpfqqszyrkp748klntdry7cmvelgw4lypqk28z8y6sy2jtrrph7r94chlmzzfupd0d One person&#39;s data pollution is another person&#39;s data gold. As long as we have next to zero insight into the immense training data sets used by LLMs this will happen again and again. Data protection fail groundhog day. #ML #AI #MLsec #security<br/><br/><a href="https://www.darkreading.com/cyberattacks-data-breaches/deepseek-breach-opens-floodgates-dark-web">https://www.darkreading.com/cyberattacks-data-breaches/deepseek-breach-opens-floodgates-dark-web</a> 2025-04-25T13:41:30Z