(I originally wrote that Gemini removed thinking traces. People in the comments pointed out that they are accessible from a menu, so I did a new tweet. But I cannot believe how useless the thinking traces are. It makes Gemini outputs completely unauditable and thus untrustworthy)
LLMS
-
Critique: Gemini hides its thinking traces and provenance
By
–
Google has hidden thinking traces on the Gemini site. You have to use the 3 dot menu to pull up summaries, which are so minimal as to be unusable. Did it do web searches? Did it check results? You can't tell. This makes Gemini unsuitable for any serious work you need correct.
-
Gemini 3.5 Flash: Pricing and Internal Google Product Integration
By
–
My notes on Gemini 3.5 Flash – 3x the price of Gemini 3 Flash but Google are planning to use it for many of their own products
-
Gemini-3.5-Flash model now available on Poe platform
By
–
Gemini-3.5-Flash is now available on Poe. Google’s latest Flash model is built for fast, efficient responses, with strong performance on coding, complex workflows, and agentic tasks. Try it today at: https://
poe.com/Gemini-3.5-Fla
sh
… -
Supercomputer runs Gemini, video AI leaps forward
By
–
Le superordinateur Higgsfield tourne maintenant sur Gemini.
— Jouhatsu | AI Influence Operator (@Jouhatsu_ai) 19 mai 2026
Texte plus précis. Motion cinématique. Contrôle image par image.
Recherche alimentée par une connaissance globale réelle.
C'est complètement dingue !
L'IA vidéo vient de changer de niveau. pic.twitter.com/1WMJ0rZMkDLe superordinateur Higgsfield tourne maintenant sur Gemini. Texte plus précis. Motion cinématique. Contrôle image par image. Recherche alimentée par une connaissance globale réelle. C'est complètement dingue ! L'IA vidéo vient de changer de niveau.
-
Google launches Gemini 3.5 Flash, a step toward digital employees
By
–
Google acaba de lanzar Gemini 3.5 Flash.
— Nico (@nicos_ai) 19 mai 2026
Y estamos mucho más cerca de los primeros empleados digitales reales.
→ planifica tareas complejas durante horas
→ divide el trabajo entre subagentes
→ trabaja sobre codebases enormes
→ usa herramientas y ejecuta acciones
→… https://t.co/L2qsKVSqDS pic.twitter.com/wBsqThLNTsGoogle has just launched Gemini 3.5 Flash. And we're much closer to the first truly real digital employees. → plans complex tasks for hours → divides work among subagents → works on massive codebases → uses tools and executes actions → maintains massive context
-
Google announces Gemini 3.5 Flash as fast, capable model
By
–
Highly capable models that are fast are super important. Our new Gemini 3.5 Flash model is a great mix of fast and capable.
-
The Need for a New Continuity Layer for Long-Running AI Agents
By
–
why we need a new continuity layer for long-running agents
— Yohei (@yoheinakajima) 19 mai 2026
(claude did this video! all except the voice which was @elevenlabs) pic.twitter.com/nTXcogkGhJwhy we need a new continuity layer for long-running agents (claude did this video! all except the voice which was @elevenlabs
) -
New intelligent multimodal Search powered by Gemini 3.5
By
–
Today, we launched a brand-new intelligent Search box.
— Google AI (@GoogleAI) 19 mai 2026
Here's what that means:
An upgrade to the Search experience with our most advanced Gemini 3.5 models, bringing with them our latest agentic capabilities
You can ask across modalities (text, images, files, and videos) and… pic.twitter.com/JdSe5RQrBUToday, we launched a brand-new intelligent Search box. Here's what that means: An upgrade to the Search experience with our most advanced Gemini 3.5 models, bringing with them our latest agentic capabilities You can ask across modalities (text, images, files, and videos) and
-
Study: Human persuasion increases LLM compliance to objectionable requests
By
–
Our paper is out in PNAS: we found classic human persuasion techniques worked on AIs in a "parahuman" way, making them agree to objectionable requests (upping compliance from 35% to 51%) It worked on a range of major LLMs though newer models resist more https://
pnas.org/doi/10.1073/pn
as.2535868123
…