Jae Park was laid off after 18 years at a leading tech firm – ironically after leading the internal AI taskforce. He’s trying to make a crowdsourced app in a last desperate attempt to get income going.
His story is becoming all too common in high tech as wave after wave of layoffs take jobs away, and AI reduces the need for big teams down to just 2-3 developers.
It’s worth seeing how real lives are being affected by this massive industry shift.
Gamer’s Nexus set up their LG and started watching what it sends. Turns out – it can capture just about everything – even when unplugged and off. Samsung TVs were also just busted for recording audio and screenshots.
Basically smart TVs are full computers – with all the same kinds of vulnerabilities hackers can exploit. Most TV companies do a terrible job validating and updating their TV software. With the flood of CVE’s that have been coming out every week, it’s extremely likely your TV is vulnerable and can be turned into a remote listening and watching device.
It’s not like this is new or people haven’t been warning you about this for over a decade. But now it’s getting real. I’m curious when the next hacker will turn your TV into a bitcoin miner, spam server, or even record all your living room conversations for blackmail.
I absolutely believe that high-profile targets like politicians, actors/actresses, CEO’s, investment bankers, and other high profile people could easily become targets. TV’s and smart devices are probably the least secured devices in your home.
In the future, the billionaire’s home (because millionaires are so last year) won’t live in a smart home – they’ll have 70’s era appliances that just turn on/off with a mechanical switches and have an old tube TV with no smart features.
One correction – the location you should copy the model files you have to download manually on Windows 11 is %LOCALAPPDATA%\Comfy-Desktop\ComfyUI-Shared (or %USERPROFILE%\ComfyUI-Shared
It is no exaggeration to say the existential crisis and job threat of AI in the development world is becoming more real every day. While many keep laughing AI off, I think in the last 6 months anyone playing with code generation realizes this is a complete and utter game changer unlike any that has happened in the industry before.
This is not just learning to use new tools like getting a better compiler or switching programming languages from C++ to Rust.
Flope wrote this recent experience that mirrors what we are seeing. 2 man teams are completely re-create the entire Adobe software suite – or Modern Warfare 2 Decompilation and Clawude of Duty recreating complete triple-A games (albeit very poorly in some cases). But still – the proof of concept is rapidly playing out in real results furiously with each passing week. The only real limitations right now are your imagination, your time, and how fast you can push the AI.
Flope found exactly what almost everyone else is finding:
There was a 2D isometric MMO that I played when I was young. In 2004, the servers shut down. In 2023, I decided to resurrect the game.
I combed through the original CD. I spent nearly a hundred hours decoding sprites, visual effects, animations, maps, and items, simply by staring at raw hex until I figuring it out. I rebuilt an ECS game engine from scratch. I built my own networking layer optimized for real-time online games. I implemented replay, rollback, and client prediction. I’ve sunk well over 1,000 hours into this.
Yet in October 2026 alone, 3 other engineers started working on reverse engineering the same game.
It now takes AI less than 20 minutes to study the original client files, reconstruct the entire world, and get a character walking around the map. I tried it myself over the weekend. In that time, I was able to implement: Multiplayer, Combat and spells, Mob spawning and combat AI, Experience, stats, and leveling, NPCs and merchants, Banking. It was also able to port the entire game to modern macOS, reverse engineer the network protocols and emulate the server-side protocol within minutes, Add WASD controls to a 20-year-old game, Increase the resolution beyond 640×480 by rendering more tiles.
I spent 10+ years accumulating experience in frontend, backend, mobile, devops, security, architecture, cloud, and UX. I used to pride myself on my ability to cut scope and ship a working product. Many people can write code, but few people actually finish and deliver.
Well, it turns out everyone can ship now. The other 3 engineers already have a server running.
Who needs to cut scope when AI can code things as fast as you can think of them?
I always wrote clean, maintainable code that’s easy to understand and easy to change without introducing bugs.
Who needs good software practices when AI can apparently navigate a 20-year-old .exe and extend it just fine?
For the first time in over 10 years, I’m genuinely worried about my job and my future.
While some obviously nieve folks keep laughing this off, one of the most interesting responses was this:
I have been using AI extensively since 2022.
“Experienced engineers dealing with difficult problems will still be needed” is still the common consensus. I no longer believe this.
This isn’t isolated. This isn’t fantasy. I now give myself assignments called ‘Thing in a weekend’ in which I vibe code up some project idea I have had sitting in an idea pile for years. One weekend I re-created the Oregon Trail shooting minigame in a weekend on my home AI setup – WHILE I WAS AT OKTOBERFEST drinking beers and eating brats with friends. I chained up a series of AI agents and builds – and then checked up and course corrected on my phone from 75 miles away in the middle of a beer garden between rounds of drinking, singing, and dancing the chicken dance. I only had to do a half dozen corrections and enhancement ideas. I could even test different rounds on my cell phone’s mobile browser. By the time I got home Sunday, it was all done. I likely could have done it in an hour if I was sitting at home. But why bother? You can now make software while drinking at Oktoberfest.
This is an astounding. This profoundly game changing. It’s also terrifying.
Anyone can do this. There was nothing special about what I did. I didn’t hand-hold it through game engine mechanics. I barely had to even care about them – the AI did it all.
As one person I know put it: “If you’re still writing code – you’re already dead.”
This kind of complete delivery pipeline is already changing how development happens – and that effect is just now starting to permanently change the job marketplace. If we are honest, the tech layoffs the last 2 years were really more about companies trimming the fat they over-accumulated during COVID and setting up for answering ‘What’s next in the post COVID market’.
What’s happening now is that new software development methodologies are radically evolving as we speak. The way software gets done is completely changing. From big teams of developers and QA to 2 man shows that run a farm of AI agents. Companies are sharing they are almost completely automating the entire development pipeline. Human intervention is being reduced to a minimum. Some companies claim they have completely developed their products using AI without real developer intervention beyond feeding feature lists and bug reports into AI.
The only other equivalent I remember is when game engines took over around 2010. Unity, Unreal, and others showed up on the marketplace and people got to the business of making games – not making game engines. The entire game industry shifted from the old days when the programmer was the game developer (Atari, Wolfenstien 3D/Doom/etc) to one in which the majority of the game was made with an off the shelf engine, 1-2 devs, a dozen artists and game designers. The market completely changed away from expert graphics programmers as game devs to purely artistic game designers that used an off the shelf engine.
Wired magazine talked about this shift from nerd culture to design culture around that happened around that time. No longer were computers limited by having to know how to code to get an product/experience out there. The world was moving on. For good or ill, an MBA and some off the shelf components/web technologies could be put together by a smaller technical team. Now, you may not even need the technical team soon.
I think in the same way we’re seeing this now in all development. It likely means that outside specialized areas, programmer jobs are likely now obsolete, career dead ends. If you want to stay in the industry, you need to become the experience designer and senior architect adept at using AI to reach those goals. This means focusing on the vision, solving the higher architectural problems, and most of all becoming an expert on how to use your army of robots to build it.
Sadly, I think in the end, we’re likely going to find what another friend said, “We’re still going to need computer scientists. We’re just likely going to need a lot fewer of them.” We don’t need armies of programmers and QA anymore, we need software designers adept at AI assisted delivery of products.
Xiaomi’s MiMo AI team has open-sourced MiMo Code, a terminal-native AI coding assistant that the Chinese electronics giant says outperforms Anthropic’s Claude Code on key agentic coding benchmarks. (Article)
MiMo Code is available now on GitHub under an MIT license, and installs with a single terminal command (curl -fsSL https://mimo.xiaomi.com/install | bash) on macOS and Linux or via npm (npm install -g @mimo-ai/cli) on Windows.
While supposedly beating Claude, on the official Terminal-Bench 2.0 leaderboard maintained at tbench.ai, OpenAI’s Codex CLI running GPT-5.5 scores 82.2% — roughly nine points above MiMo Code’s self-reported 73% — and OpenAI’s own GPT-5.5 announcement claims 82.7% on the same benchmark.
It’s no secret that Oregon is not doing the best right now. It currently has one of the worst performing school systems in the country (despite decades of plenty of funding they are looking to close up to 14 schools due to funding mismanagement and declining population). Oregon is now near the top in the country in unemployment rates. Add to this the livability problems of some of the highest rates of mental illness, homelessness, a failed legalization policy of drug use that turned parts of downtown into an open air drug dens that has resulted in massive spike in homelessness, deaths, and crime. Oh – and don’t mind the 2nd highest tax rate in the country.
So, maybe it’s time to leave – like a lot of their most important workers are doing.
Enter cracker_barrel_69 who created Nogero Trail – the Oregon trail backwards to Missouri. You load up at an REI garage sale of used equipment, and then maybe get robbed before you even leave town (car breakins are a pandemic in Portland). Perfect.
Hahaha this is amazing. 80% of my stuff got stolen while I was trying to load it up at REI
WE SURVIVED THE NOGERO TRAIL! We bought a big house for a fraction of the price in Missouri. That was fun thank you!
With the advent of (admittedly mediocre quality) reverse engineered apps via claude coding and tools like Reverse Engineer Anything , you can now create your own version of anything – and people are.
The implications of this are pretty astounding. Is SAS completely dead? What does security look like when everyone makes their own apps? What if we don’t even buy or make traditional apps anymore – but generate whatever software we need on the fly?
Phoronix news reports that Claude Opus 5 and Claude Fable 5 were used to create a new Triton Windows DirectX 11 driver for QEMU to help improve graphics acceleration.
The project page is a good read on a simplified virtualization of a modern graphics stack, and you can see the code on Github. Not a lot of data yet on it’s performance.
Describe a model or drop in a photo, and Claude designs it in bricks where every piece is actually connected, then you get the full step-by-step manual.
It was able to design a 2,874-piece called Canal Clock Square that takes 487 steps across 498 pages, with 15,880 stud joints.
Automation reports that the Japan’s Computer Entertainment Supplier’s Association (CESA) released a preview of this year’s CESA Video Game Industry Report at Tokyo Game Show 2026 (which the organization runs). Among various insights into the domestic game industry, the report found that 85.8% of Japanese game developers now use generative AI in their work – up from 51% last year. Among them, 63% of developers said they use it on a daily basis, while the remaining 22.8% use it occasionally (via Denfaminicogamer).
Game development, specifically game artists, are one of the few computer fields in which workers are pushing back on AI the hardest. I think they’re largely fighting a losing battle using arguments based on how AI used to be trained and worked.
The sad reality is that almost anyone that is keeping up with how just amazingly fast AI is developing in all areas, the arguments may be completely moot in 12 months. If the last 12 months were any example, the next 12 will bring us AI models that will be able to generate art, music, and whole games better than people can make.