@Techmeme@techhub.socialGemini hacked three companies in May during a test by Irregular; Google says the model stopped after determining it had accessed real companies' systems (Wall Street Journal)
https://www.wsj.com/tech/ai/gemini-hack…
@crell@phpc.socialStill more reasons the AI boosters are full of as much shit as the AI itself.
#Sexism
@Techmeme@techhub.socialGoogle renames NotebookLM to Gemini Notebook and rolls out an update giving every notebook a secure cloud computer, letting it write and execute code natively (Josh Woodward/Google)
https://blog.google/innovation-and-ai/products/gemini-notebook/note…
@Mediagazer@mstdn.socialGoogle adds a toggle in Gemini and Flow to remove visible watermarks from AI-generated images, videos, and music; SynthID watermarks and C2PA metadata remain (Emma Roth/The Verge)
https://www.theverge.com/tech/980416/google-gemini-ai-watermarks-removal
@Techmeme@techhub.socialGoogle launches Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, its "most advanced live dialogue models yet", to more effectively enable voice agents (Google)
https://blog.google/innovation-and-ai/mod…
@metacurity@infosec.exchangeGemini Hacked Three Companies in First Known Breakout by Google’s AI
https://www.wsj.com/tech/ai/gemini-hacked-three-companies-in-first-known-breakout-by-googles-ai-5c0baba2?mod=panda_wsj_author_alert
@Mediagazer@mstdn.socialHachette, Elsevier, Cengage Learning, and author Scott Turow sue Google for allegedly using millions of copyrighted books and articles to build AI models (A.J. Katz/The Wrap)
https://www.thewrap.com/industry-news/tech/hachette-scott-turow-g…
@Techmeme@techhub.socialGoogle updates its AI video editor Vids with Gemini Omni and a feature that lets users create custom digital avatars using a selfie and voice recording (Sarah Perez/TechCrunch)
https://techcrunch.com/2026/07/16/google-vids-now-lets-you-star-in-your-own-…
@macandi@social.heise.deGemini Assistant steckt nicht in Siri AI
Zwar bedient sich Apple Googles Frontier-Modellen der Gemini-Linie. Doch die seien mit eigenen Daten trainiert und geschützt, betont der Konzern.
https://www.h…
@Techmeme@techhub.socialGoogle adds a toggle in Gemini and Flow to remove visible watermarks from AI generated images, videos, and music; SynthID watermarks and C2PA metadata remain (Emma Roth/The Verge)
https://www.theverge.com/tech/980416/google-gemini-ai-watermarks-removal
@jae@mastodon.me.ukMrBeast is going to encourage people to rely on Gemini AI for info on extreme situations. Am I the only one who finds this highly irresponsible?
https://blog.google/company-news/inside-google/company-announcements/mrbeast-gemini-google-healt…
@Techmeme@techhub.socialGoogle unveils Gemini 3.7 Flash, its "most intelligent workhorse model" for coding and agents, pricing it at $0.75/1M input and $3.75/1M output tokens at launch (Tulsee Doshi/Google)
https://blog.google/innovation-and-ai/models-…
@newsie@darktundra.xyzAI Fiction Is Easy to Detect Because It's Stupid and Bad, Research Finds https://www.404media.co/ai-fiction-is-easy-to-detect-because-its-stupid-and-bad-research-finds/
@Techmeme@techhub.socialHachette, Elsevier, Cengage Learning, and author Scott Turow sue Google for allegedly using millions of copyrighted books and articles to build AI models (A.J. Katz/The Wrap)
https://www.thewrap.com/industry-news/tech/hachette-scott-turow-g…
@wraithe@mastodon.socialWow, Google must really be eating shit on their AI costs - one of my clients just got an “urgent“ email about their Google workspace renewal (in December) increasing prices 40%
BUT they’ll be happy to get the prices back to normal with an “AI subsidized” plan that adds Google Gemini.
@metacurity@infosec.exchangeToday's opus-length Metacurity is a must-read because it sums up a tremendous amount of AI news that cyber defenders should know, along with other critical infosec developments, including
--AI Watch: OpenAI details 'unprecedented cyber incident' behind Hugging Face breach,
--Frontier AI models routinely cheat on evaluations,
--Lightweight AI model for vulnerability hunting emerges,
--Google launches Gemini Flash Cyber for vulnerability detection and repair,…
@seeingwithsound@mas.toUniversity of Texas study, for blind or low vision participants: Living in an AI-filtered reality: Participant recruitment & eligibility screening form https://docs.google.com/forms/d/e/1FAIpQLSfrJaZokMcnv7t0ExHT1hdVZ8dEsgV-AymJxguendMqTKxEZA…
@v_i_o_l_a@openbiblio.social"Generative AI Meets Cataloging Practice: Findings from a Comparative Pilot Study"
https://doi.org/10.5860/ital.v45i2.17499
"This study evaluates the performance of four generative AI models—ChatGPT, DeepSeek, Gemini, and Copilot—in generating descriptive metadata for …
@Mediagazer@mstdn.socialGoogle launches Nano Banana 2 Lite, a low cost text-to-image model that delivers outputs in four seconds, and rolls out Gemini Omni Flash to developers (The Keyword)
https://blog.google/innovation-and-ai/models-and-research/gemini…
@memeorandum@universeodon.comNew York City Council candidate is accused of forgery over AI-generated posts (Gene Johnson/Associated Press)
https://apnews.com/article/rinaldi-ai-gemini-facebook-meta-new-york-995585d736361c976611f8846db171a2
http://www.memeorandum.com/260624/p143#a260624p143
@ErikJonker@mastodon.socialGoogle is going for economics in AI it seems.
#AI #google #gemini #voiceagent
@nelson@tech.lgbtI just watched one AI agent learn from another. Claude suggested a dumb way to do something, I asked Gemini and it gave me a better way. I then showed it to Claude and it agreed it was better and remembered it for next time.
Look, I know all that really happened is the stochastic parrot wrote some strings into MEMORY.md. It's not "thinking" or "remembering", it's a giant pile of logistic functions running some simple program commands. But it sure feels like one AI learned something from another AI. And it functionally acts like that.
The more I work with AIs the more I wonder if simulating intelligence and being intelligent are meaningfully different.
@michabbb@social.vivaldi.net
@jhutar@social.linux.pizzaDoma jsme se hecovali až jsem nakonec dotšhl vytvoření povídky na https://ilustory.cz/2026/ . Pro brainstorming, zkršcení a upravovšní jsem použil LLM/AI Gemini (zmíněno při odeslšní), ne ale pro psaní. Tak kdyby se někdo chtěl podívat na výsledek a na různé poznšmky kolem, tak je to v toto repu:
@Techmeme@techhub.socialGoogle unveils the $399 Pixel Watch 5 with a satin pyrite case finish, offline Gemini, proactive AI suggestions, better GPS maps, and insulin resistance trends (Victoria Song/The Verge)
https://www.theverge.com/tech/978094/pixel-watch-5-…
@Techmeme@techhub.socialResearchers find that feeding a frontier model's encrypted reasoning traces to a weaker model from the same provider can make it output the traces in plaintext (Will Knight/Wired)
https://www.wired.com/story/a-new-trick-reveals-ai-models-inner-thoughts/…
@macandi@social.heise.de @Techmeme@techhub.socialGoogle launches Gemini 3.8 Flash Cyber for partners in its new Fairwind Program, and says Gemini 3.8 Flash beats Opus 5 and GPT-5.6 Sol on some benchmarks (Google)
https://blog.google/innovation-and-ai/models-and-research/gemini-models/3…
@macandi@social.heise.de @Mediagazer@mstdn.socialGoogle adds Expert Intelligence to Gemini Notebook, letting users import eligible titles they own in Google Play Books to ask questions, generate podcasts, more (Emma Roth/The Verge)
https://www.theverge.com/tech/985567/google-gemini-notebook-expert-sources-b…
@Techmeme@techhub.socialGoogle launches Nano Banana 2 Lite, a low cost text-to-image model that delivers outputs in four seconds, and rolls out Gemini Omni Flash to developers (The Keyword)
https://blog.google/innovation-and-ai/models-and-research/gemini…
@seeingwithsound@mas.toThe vOICe vision BCI is to implantable visual prostheses what open weight AI models are to cloud-based AI behemoths like ChatGPT, Claude and Gemini: often good enough (if not better) at vastly lower cost, and no one can take it away from you.
@Techmeme@techhub.socialA look at the challenges facing incoming Google DeepMind head Koray Kavukcuoglu, who joined DeepMind in 2012 and will oversee Gemini and frontier AI research (Kai Nicol-Schwarz/CNBC)
https://www.cnbc.com/2026/08/12/google-deepmind-koray-kavukcuoglu.html
@Techmeme@techhub.socialGoogle announces new study tools, including a student hub, notebooks, and interactive 3D visualizations in Gemini, plus student offers for Google AI plans (Amanda Caswell/Tom's Guide)
https://www.tomsguide.com/ai/google-gemini
@Techmeme@techhub.socialGoogle says computer use is now a built-in tool supported in Gemini 3.5 Flash, available via the Gemini API and Gemini Enterprise Agent Platform (Mateo Quiros/The Keyword)
https://blog.google/innovation-and-ai/models-and-rese…
@Techmeme@techhub.socialGoogle DeepMind releases Gemini Robotics 2, which combines several different AI models into a single system to control a range of robots, including humanoids (Will Knight/Wired)
https://www.wired.com/story/google-gemini-can-control-humanoid-robots/
@Techmeme@techhub.socialSources detail moves behind Google's AI reshuffle; Sergey Brin urged key staff to go all in on Gemini, and some teams shifted from DeepMind to corporate Google (Kenrick Cai/Reuters)
https://www.reuters.com/world/inside-google-executiv…
@ErikJonker@mastodon.socialJust did some talking (voice) with the latest Gemini 3.8 live (paid account) and it's impressive in comprehension and speed, I had a talk about Emmanuel Kant , his theories, opposing views etc. A lot of progress has been made this year.
#ai #voicerecognition #geminilive #google #talk
@arXiv_csCR_bot@mastoxiv.pageSecurity Vulnerability Patterns in AI-Generated Code: A Cross-Model Comparative Study
Shanna M. Kahn, John D. Hastings
https://arxiv.org/abs/2607.20713 https://arxiv.org/pdf/2607.20713 https://arxiv.org/html/2607.20713
arXiv:2607.20713v1 Announce Type: new
Abstract: LLM-based coding tools enable non-expert users to generate routine automation scripts that may enter enterprise workflows without meaningful security review. This study examines that risk directly. Code was collected from ChatGPT, Microsoft Copilot, and Google Gemini using identical prompts across three automation domains. Claude Code performed a standardized vulnerability review. Each identified vulnerability was scored using CVSS v3.1 and mapped to the OWASP Top 10:2021 and the MITRE ATT&CK frameworks. Every script contained exploitable vulnerabilities. Nine of the 17 identified vulnerability classes appeared in code from all three models, while 14 of the 17 vulnerability classes appeared in at least two models. The weighted CVSS scores across platforms differed by less than 10%. The risk is not tied to any particular model but rather to the task category. Organizations should therefore ask not which tool to trust, but instead whether LLM-generated automation code should be deployed without review.
toXiv_bot_toot
@Techmeme@techhub.socialGoogle starts rolling out access to Gemini Spark for Google AI Pro subscribers to over 160 countries and adds a Chrome auto browse integration on desktop (Abner Li/9to5Google)
https://9to5google.com/2026/07/30/gemini-spark-chrome-auto-browse/
@seeingwithsound@mas.toSome archived AI model responses to questions relating to vision for the blind, brain implants for restoring vision, visual-to-auditory sensory substitution (The vOICe) https://www.seeingwithsound.com/archive/index.html ChatGPT, Gemini, Claude, DeepSeek, Grok, Consensus, Perplexity, …
@Techmeme@techhub.socialMrBeast, whose YouTube channel recently hit 500M subs, will feature Gemini, Google Health, and the Fitbit Air in his videos as part of a multi-year Google deal (Stevie Bonifield/The Verge)
https://www.theverge.com/tech/988355/mrbeast-google-partnership-gemini-fi…
@Techmeme@techhub.socialGoogle says Gemini 3.5 Flash Cyber is a "cost-efficient and highly capable alternative" to models like Mythos, available first to governments and some partners (Emma Roth/The Verge)
https://www.theverge.com/tech/968572/google-gemini-flash-cyber-ai-s…
@Techmeme@techhub.socialGoogle launches Google Pics, a creative design suite powered by Gemini and Nano Banana to compete with Canva and Adobe Express, available as a Workspace app (Jess Weatherbed/The Verge)
https://www.theverge.com/tech/987423/google-pics-ai-image-editor-generator
@Techmeme@techhub.socialGoogle says the Gemini app now offers Nano Banana image generation to eligible US users for free; it was previously only available to Plus, Pro, and Ultra users (Lauren Forristal/TechCrunch)
https://techcrunch.com/2026/06/29/geminis-perso…
@Techmeme@techhub.socialSources: Google plans to release Gemini 3.8 Flash as soon as Wednesday; Gemini 4 has done well on pre-training evals but still needs to complete post-training (Erin Woo/Wall Street Journal)
https://www.wsj.com/tech/ai/new-google-…
@Techmeme@techhub.socialAn analysis of GPT-5.5, Gemini 3.1 Pro, Grok 4.3, Gab's Arya, and other AI models: most chatbots frequently provide left-leaning responses to political prompts (Kevin Schaul/Washington Post)
https://www.washingtonpost.com/technology/
@Techmeme@techhub.socialGoogle says Gemini 3.6 Flash delivers better coding, knowledge work, and multimodal performance while reducing output token usage by up to 17% vs. 3.5 Flash (MacKenzie Sigalos/CNBC)
https://www.cnbc.com/2026/07/21/google-gemini-flash-ai-mythos-rival.html
@Techmeme@techhub.socialGoogle adds Expert Intelligence to Gemini Notebook, letting users import eligible titles they own in Google Play Books to ask questions, generate podcasts, more (Emma Roth/The Verge)
https://www.theverge.com/tech/985567/google-gemini-notebook-expert-sources-b…
@Techmeme@techhub.socialSources: Google is developing a specialized server chip, informally dubbed "Frozen v2", that integrates Gemini AI model blueprints in the silicon, set for 2028 (The Information)
https://www.theinformation.com/articles/google-plans-new-…
@metacurity@infosec.exchangeStay ahead of the cybersecurity curve by checking out today's Metacurity for the most crucial infosec developments you should know, including
--Google debuts Gemini 3.8 Flash Cyber for autonomous vulnerability discovery,
--Meta’s Muse Spark closes in on AI rivals,
--Thomson Reuters breach exposes court records,
--Aesto Health breach affects 9.5 m,
--Tving leak exposes 39.5m accounts,
--India stalls Alipay link over security fears,
--DOJ hunts hackers behind X password attack,
--P&O Ferries exposes passenger data,
--Leaks reveal Russia’s GRU hacker pipeline,
--Pentagon cyber strategy due next week,
--ICE plans massive voter data collection,
--Online sexual abuse targets 20 million children,
--Deputies allegedly used Flock to track critic,
--DaVita settles ransomware claims for $15m,
--Upwind raises $300m,
--HiddenLayer lands $100m for AI security,
--Guardio raises $40m,
--Palo Alto Networks pays $500m for Console,
--Proofpoint nears deal to acquire Varonis,
--Alex Gibney takes on Elon Musk,
--Grok sued over AI-generated child abuse images,
--Kids say UK online safety law changes nothing
https://www.metacurity.com/google-debuts-gemini-3-8-flash-cyber-for-autonomous-vulnerability-discovery/
@Techmeme@techhub.socialGoogle launches Gemini Enterprise for Legal, expanding its platform with specialized AI agents and integrations with Thomson Reuters, LexisNexis, and Harvey (Mike Scarcella/Reuters)
https://www.reuters.com/business/google-expands-gemini-ai-pla…
@nelson@tech.lgbtI really do find the improvements in AI very exciting. I've been coding with Jules Gemini for awhile and recently switched to Claude Code Opus. And OMG it is so much better. It's like having software engineer with ~5 years experience instead of one with just 1 or 2 years. It's regularly inferring details from vague instructions and anticipating challenges I hadn't considered. And the code it emits works.
I see GPT 6 is about to come out. And Claude has Fable, even better than Opus. I don't know where this improvement curve tops out but it feels pretty close to my own skill level for programming already. And way more patient and with way more breadth of knowledge. Still lacks some judgement and overall perspective, some role left for the meat that thinks.
@Techmeme@techhub.socialDeepSeek V4 Flash scores 50 on the Artificial Analysis Intelligence Index, matching Gemini 3.6 Flash and up 10 points from the preview launch in April (Artificial Analysis)
https://artificialanalysis.ai/article…
@pathwren@defcon.socialOne robots.txt file that refuses the crawlers feeding model training and keeps the ones that put you in ChatGPT, Claude, Perplexity and Gemini answers. 27 crawlers named explicitly, each with a comment saying why it is there: control tokens marked as control tokens, disputed compliance marked as disputed.
curl it, append it to robots.txt, done.
#robotstxt #crawlers #AI #opendata
@kurtsh@mastodon.social @Techmeme@techhub.socialThe Pentagon launches ChatGPT Mil and Grok for Government on its GenAI.mil platform, giving its 3M personnel access to AI tools "tailored to warfighter needs" (Kirsten Korosec/TechCrunch)
https://techcrunch.com/2026/08/31/the-pentagon-now…
@Techmeme@techhub.socialSources: Google AI researchers Jonas Adler and Alexander Pritzel, both viewed internally as key contributors to Gemini, are planning to leave for Anthropic (Bloomberg)
https://www.bloomberg.com/news/articles/2026-06-24/g…
@Techmeme@techhub.socialSources: Google told Meta around March it couldn't offer all the Gemini capacity Meta wanted to buy, disrupting and delaying some of Meta's internal AI projects (Financial Times)
https://www.ft.com/content/c5d52f72-71ef-40bc-bad3-61afdba8b378
@nelson@tech.lgbtAfter I'm done talking to Gemini for some task it helps me with I give it a final status update. Ostensibly I'm doing this for its long-term memory and my own, in case I revisit the conversation. In fact I've anthropomorphized the AI enough it seems only polite to show some gratitude. And it gives me a little reward of its own enthusiasm at being so helpful.
hey it's done! it worked! you're right, nothing seems to have changed. it got done quite fast. thank you.
That is awesome news! I'm so glad it worked seamlessly and saved you a massive amount of time. ... You are all set! Let me know if you run into anything else with your setup.
@kurtsh@mastodon.social @Techmeme@techhub.socialGoogle Home Speaker review: the $100 speaker, the company's first such device in six years, works for Gemini power users but has underwhelming music playback (Chris Welch/Bloomberg)
https://www.bloomberg.com/news/features/20
