<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Ayush Dixit on Medium]]></title>
        <description><![CDATA[Stories by Ayush Dixit on Medium]]></description>
        <link>https://medium.com/@ayush_dixit?source=rss-fc62085f41d4------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*n10aZy7qb8R1Wd_Ei9Y6iA.png</url>
            <title>Stories by Ayush Dixit on Medium</title>
            <link>https://medium.com/@ayush_dixit?source=rss-fc62085f41d4------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 23:23:09 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@ayush_dixit/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[AI Is Having Its 1999 Dot-Com Moment(Just Not the Way You Think)]]></title>
            <link>https://levelup.gitconnected.com/ai-is-having-its-1999-dot-com-moment-just-not-the-way-you-think-eb61218db015?source=rss-fc62085f41d4------2</link>
            <guid isPermaLink="false">https://medium.com/p/eb61218db015</guid>
            <category><![CDATA[software-engineering]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[embedded-systems]]></category>
            <dc:creator><![CDATA[Ayush Dixit]]></dc:creator>
            <pubDate>Sun, 01 Nov 2026 11:03:48 GMT</pubDate>
            <atom:updated>2026-11-02T04:44:49.628Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0xHQWf1rODjRsmzZY1ufyA.png" /></figure><p>The tourist startups are dying off, the cloud bills are coming due, and the quiet rise of Edge AI is about to take over.</p><p>Everyone is waiting for the loud, cinematic crash.</p><p>Every week, a new headline promises that the AI bubble is finally days away from bursting. Look at the numbers, and it is easy to see why investors are sweating. Worldwide AI spending has surged past $2.52 trillion, and the revenue gap: the amount of actual cash companies need to earn just to pay off their massive data center and GPU bills, has ballooned to an eye-watering $1.5 trillion!</p><p>Yet the stock market hasn’t evaporated, and the technology hasn’t vanished.</p><p>That’s because most commentators are looking at AI through a purely financial lens, comparing it to the 1999 dot-com crash. When you look at the industry from a systems and hardware perspective, a completely different picture emerges. We aren’t heading toward an extinction event. We are living through a massive, painful <strong>architectural correction</strong>.</p><h3><strong>The “Wrapper” Apocalypse is Already Here</strong></h3><p>To understand what is actually breaking, you have to separate the technology from the marketing.</p><p>For the last few years, slapping “AI” onto a pitch deck automatically inflated a startup’s valuation by 40%. Thousands of companies launched with zero proprietary technology , just a slick user interface wrapped around someone else’s giant cloud model.</p><p>That bubble <em>is</em> bursting. Enterprise buyers and everyday users have developed “subscription fatigue.” Nobody wants to pay $20 a month for ten different niche tools when the underlying tech is becoming a free commodity. Investors are finally asking the boring, brutal questions: <em>What does it cost to serve each customer? Where are the margins? Why does this need to exist?</em></p><blockquote><strong>If a company’s entire business model relies on renting someone else’s cloud servers to summarize emails, they are already dead walking.</strong></blockquote><h4><strong>The Dirty Secret of the Cloud: The Physics Problem</strong></h4><p>Here is what the software hype machine rarely talks about: <strong>physics and electricity.</strong></p><p>Training and running trillion-parameter models in massive, centralized data centers is an energy nightmare. Power grids are straining. Cooling bills are astronomical. Every time you ask a cloud-based chatbot to rewrite a grocery list, a server farm somewhere burns through compute resources that cost real money.</p><p>In engineering, when a system becomes too bloated, power-hungry, and expensive to scale, you don’t abandon the system instead you optimize it and move it closer to the metal.</p><h4><strong>The Real Pivot: From Giant Brains to “Edge” Intelligence</strong></h4><p>While Wall Street obsesses over whether tech giants can justify their trillion-dollar server farms, the smartest work in tech has shifted in the exact opposite direction.</p><blockquote><strong>The assumption that “bigger is always better” is dead. Businesses are realizing they don’t need a god-like supercomputer that knows 17th-century French poetry just to automate customer service or route warehouse inventory.</strong></blockquote><p>Instead, two massive shifts are taking over:</p><ul><li><strong>Small Language Models (SLMs):</strong> Lean, hyper-focused models that run at a fraction of the cost, offer total data privacy, and don’t require a nuclear power plant to operate.</li><li><strong>Edge AI:</strong> Moving intelligence out of the cloud and directly onto local silicon : smartphones, industrial sensors, medical devices, cars, and robotics.</li></ul><p>When AI runs locally on the device itself, latency drops to zero, cloud server bills vanish, and privacy issues disappear. The future isn’t one giant brain in the sky charging you a monthly toll; it’s millions of tiny, invisible brains embedded in the physical world around us.</p><h4><strong>What Happens When the Dust Settles?</strong></h4><p>If you are a builder, a tech worker, or someone just trying to figure out where the economy is heading, stop watching the stock tickers of overhyped software startups.</p><p>A market correction doesn’t destroy real technology; it burns away the noise. When the dot-com bubble burst, Pets.com died, but the fiber-optic cables laid in the ground stayed and gave birth to modern cloud computing.</p><p>The exact same thing is happening right now. The flashy, overvalued AI hype machine is hitting a wall of financial and physical reality. Good. Once the tourist startups clear out, the real work begins: making intelligence cheap, local, reliable, and invisible.</p><h4>Where the Jobs Are Actually Going</h4><p>While pure cloud-wrapper developers are facing a brutal hiring freeze, this architectural pivot is triggering a massive talent land grab for engineers who actually understand hardware constraints. For years, firmware and embedded systems developers worked behind the scenes while cloud and AI prompt engineers took the spotlight(and the inflated salaries).</p><p>Now, the tables are turning. You cannot run a neural network on a coin-cell battery or a constrained microcontroller using bloated Python scripts and wishful thinking. Companies are desperately hunting for <strong>Embedded ML Engineers, TinyML Architects, and Edge AI Firmware Developers </strong>builders who can take a bulky model, crush it down via quantization and pruning, and run it in C, C++, or Rust on bare metal, RTOS, or custom NPUs (Neural Processing Units) without melting the silicon.</p><p>This convergence of machine learning and low-level systems is birthing entirely new, high-paying domains. We are seeing an explosion of roles in <strong>Physical AI and Embodied Robotics</strong>, <strong>Software-Defined Vehicles (SDVs)</strong>, <strong>on-device medical diagnostics</strong>, and <strong>zero-latency industrial IoT</strong> where sensors detect factory defects in microseconds without ever pinging a server.</p><p>Even specialized niches like <strong>Edge MLOps</strong> (securely updating and managing millions of tiny models in the field) and <strong>Hardware-Aware Neural Architecture Search</strong> are becoming foundational career tracks. The AI bubble isn’t destroying tech jobs; it is shifting the premium back to true engineering. In the next decade, the highest-leverage developers won’t be the ones calling an API in the cloud they will be the ones writing the firmware that makes the physical world wake up</p><p><em>I write about embedded systems, memory architectures, and firmware engineering drawn from my work on DDR/LPDDR subsystems. If this was useful, follow along.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=eb61218db015" width="1" height="1" alt=""><hr><p><a href="https://levelup.gitconnected.com/ai-is-having-its-1999-dot-com-moment-just-not-the-way-you-think-eb61218db015">AI Is Having Its 1999 Dot-Com Moment(Just Not the Way You Think)</a> was originally published in <a href="https://levelup.gitconnected.com/">Level Up Coding</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Understanding Intel Architecture for Today’s System Firmware]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://medium.com/codetodeploy/understanding-intel-architecture-for-todays-system-firmware-f0e7a673a187?source=rss-fc62085f41d4------2"><img src="https://cdn-images-1.medium.com/max/1717/1*AdJK1fyH6ZFpyXDSrIDK2Q.png" width="1717"></a></p><p class="medium-feed-snippet">Most people have no clue what system firmware is. Say BIOS and they picture a blue setup screen. Say UEFI and they picture the same screen&#x2026;</p><p class="medium-feed-link"><a href="https://medium.com/codetodeploy/understanding-intel-architecture-for-todays-system-firmware-f0e7a673a187?source=rss-fc62085f41d4------2">Continue reading on CodeToDeploy »</a></p></div>]]></description>
            <link>https://medium.com/codetodeploy/understanding-intel-architecture-for-todays-system-firmware-f0e7a673a187?source=rss-fc62085f41d4------2</link>
            <guid isPermaLink="false">https://medium.com/p/f0e7a673a187</guid>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[software-engineering]]></category>
            <category><![CDATA[computer-architecture]]></category>
            <category><![CDATA[hardware]]></category>
            <category><![CDATA[firmware]]></category>
            <dc:creator><![CDATA[Ayush Dixit]]></dc:creator>
            <pubDate>Tue, 01 Sep 2026 13:26:28 GMT</pubDate>
            <atom:updated>2026-09-03T08:26:36.091Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[Your $1,000 Phone Trusts 32 KB of Code Nobody Can Fix]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://medium.com/codetodeploy/your-1-000-phone-trusts-32-kb-of-code-nobody-can-fix-dd2ef84efecb?source=rss-fc62085f41d4------2"><img src="https://cdn-images-1.medium.com/max/1536/1*XkLF4Wc5HGx01ADKW6-Igg.png" width="1536"></a></p><p class="medium-feed-snippet">Software engineers get patches. We get one shot, burned into silicon before the first bug report exists.</p><p class="medium-feed-link"><a href="https://medium.com/codetodeploy/your-1-000-phone-trusts-32-kb-of-code-nobody-can-fix-dd2ef84efecb?source=rss-fc62085f41d4------2">Continue reading on CodeToDeploy »</a></p></div>]]></description>
            <link>https://medium.com/codetodeploy/your-1-000-phone-trusts-32-kb-of-code-nobody-can-fix-dd2ef84efecb?source=rss-fc62085f41d4------2</link>
            <guid isPermaLink="false">https://medium.com/p/dd2ef84efecb</guid>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[firmware]]></category>
            <category><![CDATA[hardware]]></category>
            <category><![CDATA[embedded-systems]]></category>
            <dc:creator><![CDATA[Ayush Dixit]]></dc:creator>
            <pubDate>Thu, 06 Aug 2026 12:54:35 GMT</pubDate>
            <atom:updated>2026-08-06T14:59:32.230Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[Why the Chip Industry Still Writes C]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://medium.com/@ayush_dixit/why-the-chip-industry-still-writes-c-031b15a4f5d3?source=rss-fc62085f41d4------2"><img src="https://cdn-images-1.medium.com/max/1536/1*Oz2XWibuppe-SVe6RYUE2Q.png" width="1536"></a></p><p class="medium-feed-snippet">The answer has nothing to do with nostalgia. It has everything to do with what the first few milliseconds of a chip&#x2019;s life actually look&#x2026;</p><p class="medium-feed-link"><a href="https://medium.com/@ayush_dixit/why-the-chip-industry-still-writes-c-031b15a4f5d3?source=rss-fc62085f41d4------2">Continue reading on Medium »</a></p></div>]]></description>
            <link>https://medium.com/@ayush_dixit/why-the-chip-industry-still-writes-c-031b15a4f5d3?source=rss-fc62085f41d4------2</link>
            <guid isPermaLink="false">https://medium.com/p/031b15a4f5d3</guid>
            <category><![CDATA[embedded-systems]]></category>
            <category><![CDATA[software-engineering]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[firmware]]></category>
            <dc:creator><![CDATA[Ayush Dixit]]></dc:creator>
            <pubDate>Sat, 25 Jul 2026 20:27:26 GMT</pubDate>
            <atom:updated>2026-07-25T20:27:26.213Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[The Humble Mailbox: The Most Underrated Tool in Memory Subsystem Firmware]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://levelup.gitconnected.com/the-humble-mailbox-the-most-underrated-tool-in-memory-subsystem-firmware-e913b9a7e473?source=rss-fc62085f41d4------2"><img src="https://cdn-images-1.medium.com/max/1672/1*c-n5vVhuq6o_Qi2mpOioFg.png" width="1672"></a></p><p class="medium-feed-snippet">Everyone talks about queues and linked lists. Nobody talks about the mechanism that actually brings your DRAM out of reset.</p><p class="medium-feed-link"><a href="https://levelup.gitconnected.com/the-humble-mailbox-the-most-underrated-tool-in-memory-subsystem-firmware-e913b9a7e473?source=rss-fc62085f41d4------2">Continue reading on Level Up Coding »</a></p></div>]]></description>
            <link>https://levelup.gitconnected.com/the-humble-mailbox-the-most-underrated-tool-in-memory-subsystem-firmware-e913b9a7e473?source=rss-fc62085f41d4------2</link>
            <guid isPermaLink="false">https://medium.com/p/e913b9a7e473</guid>
            <category><![CDATA[computer-science]]></category>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[firmware]]></category>
            <category><![CDATA[embedded-systems]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[Ayush Dixit]]></dc:creator>
            <pubDate>Sat, 18 Jul 2026 21:48:25 GMT</pubDate>
            <atom:updated>2026-07-20T14:37:10.631Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[Mastering False Sharing: The Hidden Performance Killer in Multi-Core Embedded Systems]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://levelup.gitconnected.com/mastering-false-sharing-the-hidden-performance-killer-in-multi-core-embedded-systems-b8bbf294cab3?source=rss-fc62085f41d4------2"><img src="https://cdn-images-1.medium.com/max/620/1*vlTdo_wMJsR7odLnH1kU7Q.png" width="620"></a></p><p class="medium-feed-snippet">I&#x2019;ve spent enough time on multi-core firmware to know that some of the worst performance bugs never show up in your logic. Your code is&#x2026;</p><p class="medium-feed-link"><a href="https://levelup.gitconnected.com/mastering-false-sharing-the-hidden-performance-killer-in-multi-core-embedded-systems-b8bbf294cab3?source=rss-fc62085f41d4------2">Continue reading on Level Up Coding »</a></p></div>]]></description>
            <link>https://levelup.gitconnected.com/mastering-false-sharing-the-hidden-performance-killer-in-multi-core-embedded-systems-b8bbf294cab3?source=rss-fc62085f41d4------2</link>
            <guid isPermaLink="false">https://medium.com/p/b8bbf294cab3</guid>
            <category><![CDATA[cache]]></category>
            <category><![CDATA[embedded-systems]]></category>
            <category><![CDATA[multithreading]]></category>
            <category><![CDATA[firmware]]></category>
            <category><![CDATA[software-engineering]]></category>
            <dc:creator><![CDATA[Ayush Dixit]]></dc:creator>
            <pubDate>Mon, 06 Jul 2026 13:55:58 GMT</pubDate>
            <atom:updated>2026-07-08T16:31:14.169Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[Why Fable 5 Became the Most Controversial AI Ever Created.]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://medium.com/@ayush_dixit/why-fable-5-became-the-most-controversial-ai-ever-created-d12622829087?source=rss-fc62085f41d4------2"><img src="https://cdn-images-1.medium.com/max/1149/1*tBc3ntlzwlxFBCegItiQxw@2x.jpeg" width="1149"></a></p><p class="medium-feed-snippet">Fable 5 was supposed to be Anthropic&#x2019;s biggest launch of the year. Instead, it might end up as Exhibit A in the case that the AI bubble is&#x2026;</p><p class="medium-feed-link"><a href="https://medium.com/@ayush_dixit/why-fable-5-became-the-most-controversial-ai-ever-created-d12622829087?source=rss-fc62085f41d4------2">Continue reading on Medium »</a></p></div>]]></description>
            <link>https://medium.com/@ayush_dixit/why-fable-5-became-the-most-controversial-ai-ever-created-d12622829087?source=rss-fc62085f41d4------2</link>
            <guid isPermaLink="false">https://medium.com/p/d12622829087</guid>
            <category><![CDATA[ai-ethics]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[claude-code]]></category>
            <category><![CDATA[tech-trends]]></category>
            <dc:creator><![CDATA[Ayush Dixit]]></dc:creator>
            <pubDate>Sun, 14 Jun 2026 02:35:00 GMT</pubDate>
            <atom:updated>2026-06-14T02:35:00.733Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[The Data Structure That Runs Your Firmware (Whether You Put It There or Not)]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://levelup.gitconnected.com/the-data-structure-that-runs-your-firmware-whether-you-put-it-there-or-not-455d78d31385?source=rss-fc62085f41d4------2"><img src="https://cdn-images-1.medium.com/max/2600/1*mCFDC1QTNSoVztPrng7Sug.png" width="2816"></a></p><p class="medium-feed-snippet">My third week at a semiconductor company, I inherited a networking SoC driver framework that had been alive for eleven years.</p><p class="medium-feed-link"><a href="https://levelup.gitconnected.com/the-data-structure-that-runs-your-firmware-whether-you-put-it-there-or-not-455d78d31385?source=rss-fc62085f41d4------2">Continue reading on Level Up Coding »</a></p></div>]]></description>
            <link>https://levelup.gitconnected.com/the-data-structure-that-runs-your-firmware-whether-you-put-it-there-or-not-455d78d31385?source=rss-fc62085f41d4------2</link>
            <guid isPermaLink="false">https://medium.com/p/455d78d31385</guid>
            <category><![CDATA[software-engineering]]></category>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[embedded-systems]]></category>
            <category><![CDATA[data-structures]]></category>
            <category><![CDATA[firmware]]></category>
            <dc:creator><![CDATA[Ayush Dixit]]></dc:creator>
            <pubDate>Tue, 09 Jun 2026 22:59:40 GMT</pubDate>
            <atom:updated>2026-06-14T23:41:06.113Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[I Built a Memory Pool From Scratch. Here’s What I Learned.]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://levelup.gitconnected.com/i-built-a-memory-pool-from-scratch-heres-what-i-learned-3da736fb7350?source=rss-fc62085f41d4------2"><img src="https://cdn-images-1.medium.com/max/1774/1*8pcpWGyGYKEVhMzumDJApw.png" width="1774"></a></p><p class="medium-feed-snippet">And why it matters even more once you bring FreeRTOS into the picture.</p><p class="medium-feed-link"><a href="https://levelup.gitconnected.com/i-built-a-memory-pool-from-scratch-heres-what-i-learned-3da736fb7350?source=rss-fc62085f41d4------2">Continue reading on Level Up Coding »</a></p></div>]]></description>
            <link>https://levelup.gitconnected.com/i-built-a-memory-pool-from-scratch-heres-what-i-learned-3da736fb7350?source=rss-fc62085f41d4------2</link>
            <guid isPermaLink="false">https://medium.com/p/3da736fb7350</guid>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[firmware]]></category>
            <category><![CDATA[software-architecture]]></category>
            <category><![CDATA[freertos]]></category>
            <category><![CDATA[hardware-engineering]]></category>
            <dc:creator><![CDATA[Ayush Dixit]]></dc:creator>
            <pubDate>Sun, 24 May 2026 19:10:23 GMT</pubDate>
            <atom:updated>2026-05-26T14:13:19.320Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[Edge AI Is Eating Embedded Systems — A Firmware Dev’s Honest Take]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://medium.com/@ayush_dixit/edge-ai-is-eating-embedded-systems-a-firmware-devs-honest-take-0cfd943935b7?source=rss-fc62085f41d4------2"><img src="https://cdn-images-1.medium.com/max/1536/1*t_8ktlGt5V9BFWfe-qUOHQ.png" width="1536"></a></p><p class="medium-feed-snippet">I work on memory firmware. Here&#x2019;s what this shift actually looks like from the inside.</p><p class="medium-feed-link"><a href="https://medium.com/@ayush_dixit/edge-ai-is-eating-embedded-systems-a-firmware-devs-honest-take-0cfd943935b7?source=rss-fc62085f41d4------2">Continue reading on Medium »</a></p></div>]]></description>
            <link>https://medium.com/@ayush_dixit/edge-ai-is-eating-embedded-systems-a-firmware-devs-honest-take-0cfd943935b7?source=rss-fc62085f41d4------2</link>
            <guid isPermaLink="false">https://medium.com/p/0cfd943935b7</guid>
            <category><![CDATA[software-engineering]]></category>
            <category><![CDATA[embedded-systems]]></category>
            <category><![CDATA[hardware-engineering]]></category>
            <category><![CDATA[firmware]]></category>
            <category><![CDATA[edge-ai]]></category>
            <dc:creator><![CDATA[Ayush Dixit]]></dc:creator>
            <pubDate>Tue, 28 Apr 2026 12:13:58 GMT</pubDate>
            <atom:updated>2026-04-28T12:13:58.874Z</atom:updated>
        </item>
    </channel>
</rss>