Fable Took My Job
An engineer says frontier models like Fable helped leadership eliminate his role without matching his judgment.
A software engineer writes that models marketed at a Fable or Mythos tier did not outperform him but were cheaper than his compensation and gave leadership confidence to cut staff ahead of a possible acquisition. He argues frontier models still lack product judgment and foresight about cost, complexity, and customer needs, even as labs market them as nearing replacement-level ability. He cites Shopify using tightly harnessed models to rewrite apps in native iOS and Android, and a reported Opus 5.5 TypeScript-to-Rust port that reproduced bugs from a Go version, while saying rival model Astra failed a familiar task. He also describes wider layoffs, offshoring, and severance among engineers in his network.
- Author says model cost and leadership confidence, not skill, ended his role
- Argues frontier models still lack product taste and long-term foresight
- Cites Shopify's machine-written native apps and an Opus 5.5 code port
- Reports recent layoffs and offshoring among peers at large tech firms
Full article915 words · extracted from cendyne.dev · click to collapse
Its true. In a manner of speaking. If LLMs had not reached Fable / Mythos tier of marketed intelligence, I'd probably still have a job with the place I used to be at. For a short while.
It wasn't that Fable was better at my job. As the last engineer, Fable was cheaper than my compensation. Evidently, it gave enough confidence to the last vestiges of leadership that they could coast until M&A.
I've been through layoffs driven by budget cuts. I had to mechanically justify "I need this person because this product I'm oversee is critical to the business and still needs ownership" against a budget ceiling. Those I trusted to build out a greenfield product had their positions eliminated. The status quo was more important to maintain than resolving our operational issues with the next generation of our infrastructure.
That first meeting crushed my soul. I heard my ears ring more violently than ever. The color drained from the world. I could barely walk straight.
And, the next day, I went to a house party with two of my reports present knowing that both of them would lose their job next week with zero notice.
I know first hand what it is like to know and later execute on a layoff.
Finally, it was my time. I've had grief several times expecting it to come earlier. When it finally came, I took it well for a few hours.
Then my heart went to over 100 bpm as the fear and doubt for my future set in. And it didn't ease off for over a week.
Do you know how hard it is to sleep when all you can hear is your own heartbeat? Do you know how hard it is to function with four effective hours of sleep when you offer your body ten? It is rough.
Several in my social circles are having difficulty too. Six or Seven in a DevOps interest group have been laid off or have received a layoff notice in the last month. Work at one of FAANG of all places and oops, you're not important now! Here's 6 months of severance, we're so nice. "Ship happens" oh right, we can't ship your stuff because of some material restriction or let you pick up a box with some dignity, so we destroyed your stuff. GLHF.
Then there's that hiring manager who interviewed me at another company who then got laid off one week later because the CEO made a substantively late "AI is changing how we develop software" article. And another who lost their whole team, despite actively interviewing people in the USA, because a top-down decision was to offshore again after AI assistance raised the capability of work for the less passionate, less experienced, and less costly workforce.
"AI took my job" is far more fun to say than "AI contributed to my role being eliminated."
The frontier models are not good enough to replace an experienced engineer. They still are not, even though the marketing by the US frontier labs (and their 'I quit because I'm scared' employees) scream of what's to come.
These models lack the taste that makes for effective product decision making. It is cheaper than ever for a product manager to say "Someone complains about this on reddit, let's implement it", while the tasteful leader says 'No, that's not what our product is. Making that means we commit to it for a long time and our customers really want {something else we already do, but not perfectly} instead.'
These models lack the foresight to predict consequences weeks, months, years down the line on decisions and commitments — in operating costs, complexity, accessibility, and affordability to the very customers the product is for. All those secondary things really matter in making a product successful and worth buying. The slop services out there omit these secondary things all the time and you can taste it.
And yet, models are evidently able to give Shopify the confidence to yeet react-native out into the ocean and reach back to native iOS and android apps, entirely machine-coded while overseen by otherwise experienced and tasteful software developers.
And, I agree with Shopify. Models are getting good enough where they can be driven step by step, with a carefully crafted harness, to reproduce and port the behaviors they've A/B tested into oblivion as humans. They can push a rock up a hill. Then up the next hill, and so on — when the technique can be measured and check-pointed.
Agents are terribly good at copying things across domains. Just look at Theo's "for the memes" typescript in rust port. It ported the ts-go code to rust. According to Theo, Opus 5.5 reproduced bugs from the go version in the rust version exactly. On the other hand, Astra (the rival to Fable), failed at something so familiar, well known, and well documented.
We're heading towards the end of the year and I still do not see how these copy-machines with "super intelligence" inside are going to replace humans. What they have done is convince those with money to spend differently on their fellow humans.
Anyway, I got something else now. I get to see screens like this all day:
The joy of programming is dying. The joy of delivering results lives on.
I'm starting to think Frontier lab employees are going crazy because they're staring at this all day.