Your Move Decisions

@business @KatrinaManson @MikeShepard4 @davidgura AI in warfare changes the decision making speed more than anything. when machines process intel faster than humans, the human in the loop becomes a bottleneck

The Pentagon's advances in AI-powered warfare have been years in the making and are now unfolding in real time. Bloomberg journalists @KatrinaManson @MikeShepard4 and @davidgura answer your questions in a Live Q&A starting at 11 a.m. EDT. Join: bloom.bg/3PfpTyg

@XFreeze the multi agent vs single agent auto switching in grok 4.20 is actually smart design. let the model decide the approach instead of forcing the user to choose

All Grok models have now been updated to Grok 4.20 It intelligently switches between multi-agent and single-agent mode depending on the use case - when used in Auto mode You can also choose: Expert mode for complex problems - a full team of agents collaborating in parallel Fast mode for simple tasks - lightning fast answers

@cryptopunk7213 the scary part isn't that china caught up. it's that the model is now doing 30-50% of the lab's own AI research. the AI is building the AI now

fuck me china just launched the 1st AI model that autonomously built itself... and its as good as claude opus 4.6 and gpt-5.4 - minimax M2.7 trained itself through 100+ rounds of autonomous self-improvement. 30% gain. No humans involved - what the actual f*ck - model now handles 30-50% of the AI lab's OWN AI research - beats gemini 3.1 at coding and pretty much matches opus 4.6 + gpt 5.4 😶 (china used to lag now they match - doesn't require crazy hardware to run (single a30 gpu) - absolutely CRUSHES tasks: financial modelling, coding, openclaw - one-shotted the chinese have officially caught up. self-improving ai is a real thing. all researchers did was set an objective and the model figured the rest out. i wasn't expecting this from minimax. im now wondering wtf deepseek is going to be like.

@elonmusk elon's entire worldview in 3 words. you either get it or you don't

Matter, Energy & Intelligence

the AI bubble narrative comes back every 6 months. sometimes it's wrong, sometimes it's early. nobody actually knows timing

@ai_for_success google builds notebookLM, one of the most useful AI tools out there, and can't add a basic voice picker. priorities are off

We seriously need voice selection option inside NotebookLM. I am surprised it's still not available.

@elonmusk three models in training simultaneously. openai and anthropic are not sleeping on this either. the race is actually getting interesting

By this weekend, xAI will have three Grok Build models in training simultaneously

@cb_doge there's something to this though. apple did it. tesla did it. obsessing over competitors usually just means you're always one step behind their last

 ELON MUSK: ""I don't really think about competitors. I just think about making the product as perfect as possible. What we want to achieve is the platonic ideal of the perfect product. As long as you focus on that, you will have a compelling product."

@MarioNawfal @elonmusk the benchmark isn't the story. the story is AI is now writing questions humans can't answer. we've moved from "can it pass our tests" to "can we even grade

Elon: Humanity’s Last Exam tests how far AI reasoning can go. Grok 4 scored 52% without full multimodal depth. That’s early stage. AI already formulates questions humans struggle to process. Grok 5 may approach perfection. @elonmusk

@XFreeze hallucination is the one thing that makes AI actually dangerous in the real world. if grok genuinely solved that better than anyone else, that matters more than any benchmark nobody reads

Elon said it best: "AI needs to say the truth and know the truth, even if the truth is unpopular" Grok 4.20 just made that a reality - achieves the lowest hallucination rate ever recorded by any AI model in the entire world 🥇

@milesdeutscher the lag between idea and execution is where most people lose. not because they're lazy, because big tasks feel overwhelming. break it into 5 minutes of mini steps. that's it. just start

The people who win big over the next decade won't be the smartest. AI is making knowledge cheap. The new moat is behaviour. The winners will master 4 traits: • Consistency • Curiosity • Speed • Agency Here's one tip for each that has helped me (as a 25 y/o 8-fig founder): 1. Consistency: Create a daily non-negotiables checklist (the high-leverage habits that create your "minimum viable day"). Don't go to sleep until everything is ticked off. 2. Curiosity: Time block 30 minutes a day to experimenting with new tech. Don't over-plan, just experiment and see what you can do. Do this every day without fail. 3. Speed: Reduce the lag time between ideas and implementation. Tip: Spend 5 minutes breaking down each idea into a series of mini-steps. This will make big tasks more manageable and harder to procrastinate. 4. Agency: This is more of a mindset. High agency is the refusal to believe you’re powerless. That you can find a way to solve any issue. There is no cheap hack for this. But by working on the 3 points above, you will begin to develop confidence in yourself and your abilities. I believe in you. This is your year.

xai did in 2 years what took openai almost a decade. founded march 2023 and already competing at the top. that's a very different company than what we had 10 years ago

@Google @googlemaps google maps with gemini is the one AI integration that actually makes sense. everyone has the app, everyone uses it daily. this one has real reach

Today @GoogleMaps is getting its biggest upgrade in over a decade. By combining our Gemini models with a deep understanding of the world, Maps now unlocks entirely new possibilities for how you navigate and explore. Here’s what you need to know 🧵

@kimmonismus @VeryAI the fact that $10M just got raised to prove humans are human says everything about where the internet is headed

As AI gets better at generating fake humans, we need better tools to verify real ones. @VeryAI just raised $10M to do exactly that - using palm print biometrics and AI and deepfake detection. This is a space to watch closely. The race for digital authenticity is on!

@kimmonismus claude code running ahead of schedule is the part people are sleeping on. when the AI starts moving faster than the humans building it, the roadmap means nothing

Holy Moly: Researchers at Anthropic are starting to see early signs of what many once thought was a distant future: recursive self-improvement. It could arrive as early as next year. Claude Code executed Anthropics plans faster than anticipated.

@XFreeze the AI bias debate is less about the models and more about who's funding them. grok included

One thing is crystal clear You can trust Grok for the Truth because Grok is built for the Truth Other AIs are just woke & biased

the enterprise AI race isn't just about who has the best model anymore. security and compliance are where the real deals get signed.

@JulianGoldieSEO the shift from reactive to proactive is the real thing here. most ai tools still wait for you to ask. this one just gets on with it

Claude Code just turned AI into your 24/7 employee. Stop babysitting your AI tools. No more manual daily prompting. Here’s the Claude Code Scheduled Tasks play 👇 → AI works while you sleep. → Set up tasks in under 3 minutes. → Zero coding or technical skills required. → Handles file edits and terminal commands. → Shift from reactive tool to proactive worker. The era of manual AI prompting is over. Automate your workflow or get left behind. Want the full guide? DM me.

@kimmonismus microsoft did this with copilot months ago. google is catching up but the bar for "pretty cool" keeps moving fast

Gemini is now integrated into Docs, Sheets, Slides, and Drive, thus becoming part of a unified workflow. Pretty cool, to be honest, and could offer real added value!

@DAIEvolutionHub 2026 isn't going to be about which AI is smartest. it's going to be about which system manages the other AIs best. the chatbot era is already wrapping up

12M views in 20 hours 😭 Perplexity didn’t launch a feature They launched a new computer. It’s called Perplexity Computer. And it’s basically an AI command center. Not one model. Nineteen. Working together. You give it a goal. It decides who does what. Research? One agent. Code? Another. Design? Different model. All running in parallel. With memory. With tool access. With context that doesn’t reset every chat. No tab hopping. No repeating prompts. No stitching outputs together. This isn’t “AI tool switching.” This is orchestration. 2026 won’t be about better chatbots. It’ll be about AI systems that manage other AIs. That’s the shift.