🚀 Welcome to AI Unraveled (Feb 18, 2026):
This episode is made possible by AIRIA. 🛑 Stop fearing Shadow AI. Orchestrate it. With the release of Claude Sonnet 4.6 and the viral rise of OpenClaw, your employees are deploying intelligence faster than your IT department can track it. AIRIA is the “Control Plane” for your entire enterprise AI stack, giving you unified security, real-time cost control, and visibility into every agent loop—whether it’s on a server or a local Raspberry Pi.
👉 Secure Your Stack: Get the Airia Demo here.
Today’s Briefing: We dive into Anthropic’s “trickle-down” strategy as Sonnet 4.6 matches flagship performance at a fraction of the cost. We also analyze Apple’s aggressive push into AI wearables, the escalating feud between Anthropic and the Pentagon, and the $200 billion investment from Micron to break the memory bottleneck. Plus: Why Raspberry Pi stock is soaring and the “ROI Regret” hitting 74% of CIOs.
You can translate the content of this page by selecting a language in the select box.
We want to share an exciting opportunity for those of you looking to advance your careers in the AI space. You know how rapidly the landscape is evolving, and finding the right fit can be a challenge. That's why I'm excited about Mercor – they're a platform specifically designed to connect top-tier AI talent with leading companies. Whether you're a data scientist, machine learning engineer, or something else entirely, Mercor can help you find your next big role. If you're ready to take the next step in your AI career, check them out through my referral link: https://work.mercor.com/?referralCode=82d5f4e3-e1a3-4064-963f-c197bb2c8db1. It's a fantastic resource, and I encourage you to explore the opportunities they have available.
Strategic Pillars & Key Topics:
🧠 The Efficiency Shift
Claude Sonnet 4.6: Anthropic’s new mid-tier model beats its own flagship Opus 4.6 in finance and coding benchmarks while costing 80% less.
Qwen-3.5: Alibaba’s new open-weight model delivers frontier performance with massive inference gains, activating only 17B out of 397B parameters.
⌚ Hardware & Wearables
Apple’s AI Eyes: Bloomberg reports Apple is fast-tracking camera-equipped glasses, pendants, and AirPods to give Siri real-time visual awareness.
Raspberry Pi Surge: Enthusiasts are betting on $RPI as the preferred platform for running OpenClaw personal agents locally.
🛡️ Defense & Security
Pentagon Drone Swarms: SpaceX and xAI enter a $100M contest to build voice-controlled autonomous drone swarms.
Anthropic Feud: The Pentagon threatens to label Anthropic a “supply chain risk” over the lab’s refusal to remove safety guardrails for military use.
Lockdown Mode: OpenAI introduces a “Lockdown Mode” for ChatGPT to prevent prompt injection and data leaks.
📉 Business & ROI
The CIO Regret: A new Dataiku report finds 74% of CIOs regret at least one major AI vendor decision made in the last 18 months.
Alpha School Controversy: Inside the $65,000 AI-powered school accused of generating faulty lesson plans and “betraying” students.
Credits: This podcast is created and produced by Etienne Noumen, Senior Software Engineer and passionate Soccer dad from Canada.
Timestamps:
00:00 – Intro: Anthropic’s pricing shock with Sonnet 4.6, Apple’s visual AI strategy, and the drone swarm battle.
01:05 – Sponsor Segment: Managing Shadow AI and orchestration with AIRIA.
02:08 – Claude Sonnet 4.6: The “trickle-down” strategy—beating the flagship Opus 4.6 for 1/5th of the price.
06:55 – Apple’s AI Eyes: The fast-tracking of smart glasses, pendants, and camera-equipped AirPods.
10:15 – Figma & Claude Code: Integrating “vibe-coded” prototypes into editable design canvases.
13:40 – Sponsor Mid-Roll: Orchestrating the “Shadow AI” movement with AIRIA.
15:52 – Defense & Geopolitics: SpaceX joins the drone swarm contest; the Pentagon-Anthropic standoff escalates.
19:15 – Hardware Boom: Micron’s $200B investment in HBM to break the AI memory bottleneck.
22:10 – OpenClaw & Raspberry Pi: The single-board computer mania and the risks of DIY personal agents.
27:30 – The ROI Crisis: Why 74% of CIOs regret their recent AI vendor decisions.
31:45 – Alpha School Controversy: Inside the AI-powered private school accused of “betraying” students with faulty lessons.
35:50 – Conclusion & Outro: The signal of “Disruption” and a final message from AIRIA.
Keywords: Claude Sonnet 4.6, Apple Smart Glasses, AI Wearables, OpenAI Lockdown Mode, Micron Memory Boom, Raspberry Pi AI, OpenClaw, Pentagon Drone Swarms, Anthropic vs Pentagon, Alpha School AI, Qwen-3.5, AIRIA, Shadow AI, CIO ROI Regret.
🚀 Reach the Architects of the AI Revolution
Want to reach 60,000+ Enterprise Architects and C-Suite leaders? Download our 2026 Media Kit and see how we simulate your product for the technical buyer: https://djamgamind.com/ai
Connect with the host Etienne Noumen: https://www.linkedin.com/in/enoumen/
⚗️ We Practice What We Preach: AI Unraveled is produced using a hybrid “Human-in-the-Loop” workflow.
Anthropic’s powerful Claude Sonnet 4.6
Image source: Anthropic
Anthropic just rolled out Claude Sonnet 4.6, its latest mid-tier model that matches or beats the flagship Opus 4.6 across finance, computer use, coding, and office benchmarks at 1/5 of the price — while featuring a 1M token context window.
The details:
-
On SWE-Bench Verified for coding, Sonnet 4.6 came in at 79.6%, just below Opus 4.6’s 80.8% — while costing just 1/5 as much to run.
-
Sonnet 4.6 outscored Opus 4.6 on agentic financial analysis and office-task benchmarks, a first for the mid-tier Claude model.
-
Early Claude Code testers preferred Sonnet 4.6 over its predecessor 70% of the time, also winning over Opus 4.5 at a 59% rate.
-
Computer use capabilities also keep climbing, with Sonnet’s OSWorld scores jumping from under 15% in late 2024 to 72.5%.
Why it matters: Anthropic is running a trickle-down playbook at warp speed, shipping near Opus-caliber capabilities to its cheaper line just weeks after the flagship upgrade. With strong Chinese models continuing to undercut everyone on price, Sonnet 4.6 looks like Anthropic’s bid to better compete for the volume layer of the agentic boom.
Apple going all-in on AI wearables
Image source: Lovart / The Rundown
Apple is reportedly fast-tracking three camera-equipped AI wearables, according to Bloomberg — including smart glasses, a pendant, and new AirPods all designed to give Siri real-time visual awareness through the iPhone.
The details:
-
The glasses will feature dual cameras, Apple-designed frames, no display, and a production target of late this year ahead of a 2027 launch.
-
The pendant will act as an always-on camera and mic for your iPhone, internally dubbed the phone’s “eyes and ears”.
-
Camera-equipped AirPods could ship as early as this year, using low-res sensors to feed Siri visual context and building on live-translation features.
-
All three devices will tie into Apple’s revamped Siri, which is expected to get a chatbot-style interface in iOS 27 later this year, powered by Google’s Gemini.
Why it matters: AI wearables from Apple (that actually feature a working model) would immediately shake up the mainstream hardware landscape and vault the tech giant into the AI spotlight. But these devices depend on a Siri overhaul that we still need to see to actually trust after years of absolutely brutal delays and underdelivering.
Figma turns Claude Code builds into editable designs
Image source: Figma
Figma just introduced a new “Code to Canvas” integration with Anthropic that lets developers capture interfaces built in Claude Code and convert them into fully editable design files on Figma’s canvas.
The details:
-
The feature grabs live UI from a browser and turns it into native Figma layers that can be duplicated, annotated, or rearranged.
-
Figma’s existing MCP server closes the loop, letting developers pull edited designs back into coding environments without losing the shared context.
-
Devs can capture entire multi-step flows at once, keeping the full user journey intact so teams can review and edit the experience side by side.
-
The launch comes as Figma stock has cratered roughly 85% from last summer’s high amid a broader SaaS selloff driven by AI coding fears.
Why it matters: This generation of AI coding tools has made it trivial to build a working UI, but Figma is hoping to be the polished, shippable design layer on top of raw vibe-coded prototypes. But with model capabilities only improving, that polishing layer may also soon be automated — something the markets appear to notice as well.
SpaceX competes in Pentagon AI drone swarm contest
-
SpaceX and its subsidiary xAI are competing in a secretive $100 million Pentagon contest to build voice-controlled, autonomous drone swarming technology, marking a new move into AI-enabled weapons development for Elon Musk.
-
The six-month competition, launched by the Defense Innovation Unit and the new Defense Autonomous Warfare Group, will progress in five phases from software development to real-life testing of drones for offensive purposes.
-
OpenAI is also involved through a partner company called Applied Intuition but limited to voice-command translation, while SpaceX and xAI are expected to work on the entire project together, raising concerns among defense officials.
Anthropic-Pentagon feud escalates over AI use
The Pentagon is reportedly “close” to cutting ties with Anthropic and designating the company a “supply chain risk” — a badge usually reserved for foreign adversaries — over its restrictions on how Claude is used by the military.
The details:
-
The designation of “supply chain risk,” if given, would force all U.S. defense contractors to cut ties with Anthropic — hitting the AI’s major business severely.
-
Defense officials are demanding the right to use AI for “all lawful purposes,” while Anthropic is holding firm against granting broad permissions.
-
The company is open to loosening restrictions but wants to ensure its AI is not used for spying on Americans or building autonomous weapons, Axios reports.
-
Claude is currently the only AI on the Pentagon’s classified systems, and was also reportedly used via Palantir to capture Nicolás Maduro in January.
Why it matters: Experts have long warned about the unchecked use of AI in warfare, and this standoff marks a notable moment showing the growing friction between companies’ responsible-use guardrails and the military’s operational demands. Only time will tell which side shapes the rules of AI in national security.
Alibaba nears frontier with open-weight Qwen-3.5
Image source: Qwen
Alibaba’s Qwen released Qwen3.5-397B-A17B, an open-weight vision language model featuring a “hybrid architecture” that delivers massive inference gains while rivaling proprietary giants like OpenAI’s GPT-5.2 and Google’s Gemini 3 Pro.
The details:
-
Qwen-3.5 uses a sparse MoE design, activating only 17B parameters out of 397B for each query, balancing high-level capabilities with low latency.
-
The model is close to frontier players across the board, and even surpasses them in domains like agentic search, doc recognition, and instruction following.
-
Alibaba claims it is 60% cheaper to use and at least 8x better at processing large workloads than its immediate predecessor, Qwen3-Max.
-
The release is aimed at handling continuous, multimodal reasoning required by agents, although it doesn’t seem very good at running a vending machine yet.
Why it matters: Chinese labs are on a roll, and with Qwen3.5 combining near frontier performance, 60% lower costs, and open weights, the race is clearly shifting toward efficiency and scalability. If the momentum continues, the AI balance may hinge less on raw size and more on who can deliver powerful models at the lowest rate.
OpenAI adds Lockdown mode to ChatGPT
-
OpenAI has added a new Lockdown Mode and an “Elevated Risk” warning label to ChatGPT, both designed to protect users from prompt injection attacks that can trick AI into leaking sensitive data.
-
Lockdown Mode is optional and deterministically limits or fully disables high-risk features like web browsing, while the “Elevated Risk” label warns users before actions like opening unverified external links.
-
Business plan users on ChatGPT Enterprise, Edu, Healthcare, and Teachers already have the protection, but OpenAI said it will roll out to other users in the coming months across unknown payment tiers.
Micron invests $200B to break the AI memory bottleneck
-
Micron Technology is spending $200 billion to expand its U.S. manufacturing, including two new fabs in Boise and a $100 billion complex near Syracuse, to address the AI memory chip shortage.
-
Demand for high-bandwidth memory chips has far outpaced supply as AI data centers have grown, with DRAM contract prices rising over 170% and DDR5 prices jumping nearly 500% since September.
-
Micron’s gross margins have climbed from 18.5% in early 2024 to 56% recently as it shifted toward data-center HBM chips, and its share price has risen more than sixfold since last April.
OpenClaw boosts interest in Raspberry Pi
Shares in the British single-board computer makers are spiking thanks to enthusiasm around OpenClaw, adding more than $1 billion to its market value so far. Bloomberg suggests that the incredible rally is all thanks to a Long $RPI social media post from Monday (which didn’t even go that viral), suggesting that the company’s SBCs will become a preferred platform for running OpenClaw cheaply. If Raspberry Pi becomes the new Mac Mini, this could turn out to be a smart bet.
Ready to install OpenClaw? Do it smart, or wait
When OpenAI hired OpenClaw founder Peter Steinberger, it officially turned personal AI agents into the hottest trend of 2026.
If you’re feeling FOMO and are about to spend $500 on a Mac mini to install OpenClaw and spin up your own personal AI assistant, there are a few factors you may want to consider first.
Despite what you may have heard, installing OpenClaw is a pretty technical and time-intensive process. There are four-hour YouTube videos that walk you through the entire process —and that doesn’t include all the prep and planning it takes to do it right by setting up separate accounts for email, texting, GitHub, Slack, an Apple account, and any other services you want your personal AI assistant to work with.
That brings us to the second caveat: You shouldn’t install this on your main computer or allow it to use your logged-in accounts for email, file storage, text messaging, etc. That would be like hiring a new employee and giving them the password to your laptop and your email on the first day.
Remember that AI agents are non-deterministic systems. That means they don’t follow a set of step-by-step instructions. In fact, they can delete all of your files because they perceive that you are well-organized and thought it would be helpful to clean up for you.
So instead of installing OpenClaw yourself right now, consider one of these options:
-
Use a cloud-based OpenClaw service like Hostinger, MyClaw.ai, or V2Cloud. You can spin up one of these almost instantly for $5-$10 per month, and they give you a secure sandbox that keeps your personal AI assistant from accessing other machines and data on your network.
-
Wait until OpenAI and Steinberger release their personal AI assistant product that will “bring agents to your mom and everyone else,” as OpenAI’s CMO Kate Rouch said. We should expect that to happen relatively quickly.
-
Give Anthropic’s Claude Cowork a try. It’s another AI agent, but it has more guardrails to keep you from getting in trouble. For enterprises, there’s also the lesser-known Amazon Quick Suite, which offers many of the benefits of personal AI assistants within the confines of a traditional IT environment.
Conflicting signals: AI investments vs. ROI doubts
AI investments continue surging, evident in headlines, fundraising rounds, stock rallies, and product launches. But executives are having second thoughts.
The AI company Dataiku released a new global report based on a Harris Poll survey of more than 800 CIOs worldwide, and it found that CIOs are not only facing regret from their AI investments, but also hold anxiety about what AI’s ability to perform means for their organization’s future and their jobs.
A majority of CIOs (74%) said their role will be at risk if their company does not deliver measurable business gains from AI within the next two years. At the same time, they are not seeing the results, yet they are being questioned about them. The report found that:
-
74% say they regret at least one major AI vendor or platform decision made in the last 18 months.
-
62% say their CEO has directly questioned or challenged those decisions.
-
Nearly one-third (29%) say they have repeatedly been asked to justify AI outcomes they could not fully explain.
“ROI is a real question, but the honest answer is that we’re early. It’s normal that measurement frameworks haven’t caught up with a technology whose application is still being defined,” Kurt Muehmel, Head of AI Strategy at Dataiku, told The Deep View. “The pressure is real. But the answer isn’t to stop investing, it’s to stop investing badly.”
To maximize the value of AI investments, Muehemel recommends avoiding a single model provider. Advantages of this approach include: switching to better models as they evolve rapidly, leveraging cheaper alternatives if the AI bubble pops, and avoiding the need to rebuild your entire system when swapping out a model.
A Gartner report that surveyed more than 300 CFOs and finance leaders also found that they are willing to increase AI spending in 2026 based on the future promise of AI. The report found that 60% of CFOs plan to increase AI investments in the finance function by 10% or more in 2026, while another 24% expect gains of 4% to 9%.
“This investment surge is driven by a ‘Return on the Future’ mindset, which prioritizes long-term strategic disruption and competitive parity over immediate financial gains,” said Nauman Raja, Director Analyst at Gartner. “After all, 88% of CFOs view AI as a critical mandate for future efficiency, so its potential of being a disruptive force is too great for CFOs not to invest in and try to achieve gains with.”
Students Are Being Treated Like Guinea Pigs:’ Inside an AI-Powered Private School
Alpha School, an “AI-powered private school” that heavily relies on AI to teach students and can cost up to $65,000 a year, is AI-generating faulty lesson plans that internal company documentation find sometimes do “more harm than good,” and scraping data from a variety of other online courses without permission to train its own AI, according to former Alpha School employees and internal company documents.
Alpha School has earned fawning coverage from Fox News and The New York Times and received praise from Linda McMahon, the Trump-appointed Secretary of Education, for using generative AI to chart the future of education. But samples of poorly constructed AI-generated lessons that I have viewed present students with unclear wording and illogical choices in multiple choice questions.
“These questions not only fail to meet SAT standards but also fall short of the quality we promise to deliver,” one employee wrote in the company’s Workflowy, a company-wide note taking app where every employee can see what other employees are working on, including their progress and thoughts on various projects. “From a student’s perspective, when answer options don’t logically fit the question, it feels like a betrayal of their effort to learn and succeed. How can we expect students to trust our assessments when the very questions meant to test their knowledge are flawed?”
What Else Happened in AI on February 18th 2026?
xAI began rolling out the long-awaited Grok 4.20 in a public beta, featuring a new agent workflow that uses four agents working in parallel to research and handle tasks.
Meta and Nvidia announced a new multiyear AI chip deal spanning millions of GPUs and CPUs to help power the tech giant’s AI infrastructure buildout.
Cohere Labs open-sourced Tiny Aya, a 3.35B parameter multilingual model that handles 70+ languages with strong gains for typically underrepresented dialects.
French AI startup Mistral made its first-ever acquisition, with serverless platform Koyeb joining to boost its Mistral Compute cloud infrastructure arm.
WordPress launched a new AI assistant capable of editing layouts, generating images, and rewriting content directly inside the editor for streamlining website design.
Meta patented a social networking system that uses AI trained on a user’s interaction data to simulate their responses when they’re on a long break, or even deceased.
India kicked off its AI Impact Summit, hosting execs from global AI giants, including OpenAI’s Sam Altman, Google’s Sundar Pichai, and Anthropic’s Dario Amodei.
Sam Altman and Dario Amodei confirmed India is now the second-largest market for ChatGPT and Claude, with Amodei also announcing Anthropic’s Bengaluru office.
Ireland’s Data Protection Commission is probing xAI’s Grok over concerns it can generate sexualized images of women and children, after similar UK and EU action.
SpaceX (and xAI) will reportedly compete in the Pentagon’s $100M contest to produce voice-controlled, autonomous drone swarming technology.
ElevenLabs launched its “ElevenLabs for Government” initiative to help public sector agencies deploy secure, multilingual voice and chat AI.
We want to share an exciting opportunity for those of you looking to advance your careers in the AI space. You know how rapidly the landscape is evolving, and finding the right fit can be a challenge. That's why I'm excited about Mercor – they're a platform specifically designed to connect top-tier AI talent with leading companies. Whether you're a data scientist, machine learning engineer, or something else entirely, Mercor can help you find your next big role. If you're ready to take the next step in your AI career, check them out through my referral link: https://work.mercor.com/?referralCode=82d5f4e3-e1a3-4064-963f-c197bb2c8db1. It's a fantastic resource, and I encourage you to explore the opportunities they have available.
