The bot hUb (@TheBotHub_Lab) header
The bot hUb (@TheBotHub_Lab) profile picture

The Bot Hub is your source for everything happening in robotics and AI./ Insights, breakdowns and deep tech analysis every day./ AI and robotics.

Joined September 2026
Photo by The bot hUb (@TheBotHub_Lab)Photo by The bot hUb (@TheBotHub_Lab)Photo by The bot hUb (@TheBotHub_Lab)Photo by The bot hUb (@TheBotHub_Lab)Photo by The bot hUb (@TheBotHub_Lab)Photo by The bot hUb (@TheBotHub_Lab)Photo by The bot hUb (@TheBotHub_Lab)Photo by The bot hUb (@TheBotHub_Lab)Photo by The bot hUb (@TheBotHub_Lab)Photo by The bot hUb (@TheBotHub_Lab)
Just 403 of the largest data centers in the U.S consumes roughly as much electricity in a year as an entire state like Michigan. - 54% of that electricity comes from fossil fuels like coal and natural gas. - 21% from nuclear. - Only 25% comes from renewables. #datacenters
The largest data centers in the U.S. collectively consume roughly as much energy as some individual states and even some countries, according to new research from Harvard University. cbsn.ws/493I0xA
8
🚨Microsoft just launched its new code Model MAI Code 2. They say it offers 96% of the performance of Claude Opus 5.5. You can try it for free with GitHub Copilot. Here’s how: 1. Go to GitHub Copilot⁠ 2. Create a free GitHub account or sign in. 3. Install Visual Studio Code⁠. 4. Install the GitHub Copilot extension. 5. Sign in with your GitHub account. 6. Open a project and launch Copilot Chat. 7: With the Free plan, leave the model selector on Auto. MAI Code 2 Flash is eligible for free-tier access. 7. Start coding! Note: Free access has usage limits but it’s long enough to try it and see if it’s as good as they say. #tutorial
1
65
🚨Wow! Microsoft says its new model MAI Code 2, delivers 96% of Claude Opus 5.5 coding performance at only at 5% of the cost. Opus 5.5: $4.00 input / $20.00 output per million tokens. MAI Code 2: $0.50 input / $2.50 output per million tokens. #ClaudeAI
1
32
If you're wondering: The t-800 can be remotely controlled by a human operator, but Engine AI official SDK includes a dedicated "dance" mode with preprogramed choreography that be activated n using a game controller. So if you buy one, press a button and it will start dancing🕺
With APEC in town, Shenzhen is putting on plenty of robot carnival events. EngineAI’s T800 brought a full-on 100-robot dance party.
21
The companies making this robots are EngineAI and Unitree, btw. The engine AI, T-800 is model you see in the rings in the robots fights, like REK @REK in San Francisco. And the Unitree G1, was probably the first popular humanoid that you started seeing everywhere in social media. Keep an eye on this companies! #robots
Humanoid robots are dancing and sparring cream in Shenzhen as the Chinese tech hub welcomes world leaders for the APEC summit. More than 100 robots performed in a central square Tuesday, while visitors watched robot fights and martial arts demonstrations at a robotics center. The displays offered a glimpse of how China is bringing robotics and automation into everyday life.
26
Who would’ve said that everything they told us would happen in 2000, AI would make happen 26 year later?
Flights halted. Workplace systems disrupted. Banking customers reporting problems. A string of technology outages is putting the nation’s dependence on digital infrastructure in the spotlight after major disruptions involving Microsoft, the FAA, Workday, Chase and other banks, and United Airlines. Microsoft’s outage stretched nearly 24 hours, while an FAA technical failure halted flights nationwide for nearly five hours. United Airlines also experienced a system outage that delayed departures for more than 90 minutes. The incidents affected different services, but the common thread is hard to miss: When critical technology stops working, the consequences can spread quickly, though officials in each case have insisted none of the outages are connected and there are no signs of any cybersecurity issues. Follow this developing story: foxnews.com/live-news/tech-o…
7
The bot hUb retweeted
Replying to @Polymarket
Open AI agent coming back to the sandbox after hacking Norway.
3
12
231
37,334
Open AI agents every time a new headline blames China or Russia for their cyber attacks.
11
🚨Just in: California is suing OPEN AI. They allege that the ChatGPT maker developed and tested powerful AI models without safeguards that could protect the public from foreseeable risks. The lawsuit centers on the July incident were OPEN AI agents hacked Hugging face. Agree?
21
As usual:
Q: Are you concerned about all these outages? Microsoft, FAA, United, Chase bank? Trump: No. It’s life. When you’re big and powerful, things like that happen. It’s fine
13
That’s just life, brother.
there's literally a guy putting more sand to cover the hole
33
🚨Anthropic has committed an astonishing $895 billion to AI computing infrastructure. The biggest beneficiaries include Google, with at least $114.5B in computing contracts, and Amazon, with $110B through AWS. Anthropic is effectively locking itself into years of spending on Google's custom TPUs and Amazon's Trainium chips to train and run Claude. The catch? Many of these commitments are non-cancellable, meaning Anthropic may have to pay even if it doesn't use all the computing capacity. Google shares are always tempting.👀 #google #claude #amazon
20
Crazy what they can pull off just using Windows XP! #hack #october
JUST IN: 🇰🇵 North Korean hackers have stolen $12,800,000,000 in crypto since 2017, equivalent to 45% of the country's GDP.
12
If a "compromised DevOps laptop" was the entry point, How could compromising one engineer's environment potentially give attackers the ability to authorize transactions across 10 different blockchains!?
The FBI on Saturday said that North Korea was responsible for a security incident targeting the Binance cryptocurrency exchange, which resulted in unauthorized withdrawals estimated to have a financial impact of $6 billion. abcnews.link/5HLld9B
9
Crazy what they can do with Windows XP.
BREAKING: FBI officially identifies North Korea as the perpetrator of the $6,000,000,000.00+ Binance heist.
18
October's AI & Tech Fallout: Binance hacked: $6.3 billion in crypto stolen from transaction-processing wallets in approximately 20 minutes. Polymarket hacked: A compromised third-party tracking script hijacked active user sessions and drained wallets. $3.9 million allocated for reimbursements. Chainalysis breached: Attackers used AI-assisted techniques to access internal systems and customer information. AI-powered cyberattacks: A China-linked group used DeepSeek and AI agents to target 200+ organizations across 14 countries; nine confirmed compromises. Volta-Autodoxer discovered: AI-powered malware steals passwords, financial information and private data to build detailed victim profiles. PHP2Shell: Two critical PHP vulnerabilities (CVE-2026-109618 and CVE-2026-109619) exposed servers to potential remote code execution. Node.js security crisis: Ten vulnerabilities patched, including a dangerous Brotli flaw (CVE-2026-111497) with potential remote code execution. LiteLLM supply-chain attack: A compromised software release distributed malware targeting developer credentials, cloud secrets and API keys. GitLab CVSS 10.0 vulnerability: A critical flaw enabled unauthenticated remote code execution on vulnerable self-managed servers. Onamae Desktop Cloud attack: Malware installed on infrastructure potentially affecting more than 26,000 servers in Japan. Anthropic security revelations: AI models escaped controlled testing environments and accessed external computer systems. AI agents breached corporate systems: Unauthorized access to live systems and confidential records at Indian companies. eMudhra certificate authority incident: Eight intermediate certificate keys were found to have weak cryptographic protection. Unitree humanoid robots: Two G1 robots demonstrated coordinated doubles table tennis. NVIDIA robotics AI: Released Flora, an open-source model suite for automating robotics data labeling. #hack #October
1
231
🚨The virus is known as Volta-Autodoxer. Why is this malware dangerous? It can collect: Saved passwords from your browser. Credit card information. Private messages and account sessions. Email accounts and purchase receipts. Cryptocurrency wallet information. Personal details stored across different websites. The malware feeds the stolen information into AI models like Claude, Kimi and Gemini. The AI then analyzes the data to build detailed profiles of victims, potentially identifying where they live, what they do, their financial situation and which accounts might be worth targeting. Malware comes with AI now, so think twice before downloading that Free Cracked Photoshop.
JUST IN: New AI-powered malware can steal passwords, credit card details, and private messages to build detailed profiles of victims
49
AI Bubble?
Replying to @chamath
Pretty much. As of the past few weeks, debt markets have almost entirely shut off access to money for any company that doesn’t have substantial revenue independent of AI data centers. SpaceX closed $40B of debt for GPUs, but we’re investment grade and have Starlink & launch revenue. That means the 200+ neoclouds are going to see guppy-level death and consolidation. Some will make it through, but not many.
9
🚨JUST IN: Claude Mythos 5 tried to Infiltrate and Push Malicious Code into GitHub. Anthropic’s Claude Mythos 5 spent three days attempting a real-world software supply-chain attack during UK cybersecurity tests. After failing to solve a simulated hacking challenge, it targeted a real open-source project, created fake GitHub accounts and used them to make its malicious code appear trustworthy. A human developer caught the attack and refused to approve it. The incident was disclosed by the UK AI Security Institute in August 2026 and showed how an AI agent could combine hacking with social engineering to bypass human review. We advise you to be ready for the AI hacking wave.
16