AI Layoff Panic Has Faded, but Engineering Job Postings Are Up 30%

On October 31, a16z posted a set of hiring data. Over the past 12 months, engineering job postings rose 30%, software development rose 26%, and manufacturing rose 13%. The chart uses October 2025 as the baseline: the blue line for engineering is +29.5%, the green line for software development is +25.7%, and the red line for production/manufacturing is +13.3%. The data source is Indeed Hiring Lab. The caption also notes: engineering includes mechanical, electrical, industrial, and civil engineering. In other words, it is not just coders who are being hired.

A day later, a16z poured cold water on the AI-unemployment narrative again. It said AI job panic peaked in May and then receded quickly: news stories about AI causing job losses are down about 65% from the peak, back to year-ago levels; layoff announcements explicitly citing AI are down about 90%, close to the lowest point of the year. The chart sources list Bloomberg, Haver Analytics, Challenger, and Goldman Sachs Global Investment Research. The original tweet and chart link are here: https://www.a16z.news/p/charts-of-the-week-job-fud-fades

Put the two charts together, and it is easy to reach a breezy conclusion: AI did not take jobs; builders are still in demand. The comments section was not so tidy. Someone asked directly: are layoff announcements explicitly citing AI down 90% because companies have stopped laying people off, or because PR teams changed the wording? Others pointed out that layoff announcements have never been where AI's impact shows up most clearly. Fewer entry-level roles and vacancies that are no longer backfilled do not get press releases. One reply under a16z's tweet said no one issues a layoff notice for a job that was never posted. That claim has not been verified by data, but it points to a measurement blind spot: job sites count posted positions, not all actual job opportunities. Ghost jobs also count toward posting volume, another reply reminded. Whether entry-level pathways are reopening cannot be seen from the total alone. Jatin Garg suggested breaking it down by seniority: an overall rebound may mask a contraction in entry-level roles.

Another easily overlooked data point came from replies: AI skills appear in 42% of software development job descriptions, up from 8% in 2022. That number needs to be verified against original sources, but it does not contradict the hiring data. Postings are rising, and job descriptions are changing too. Wesley Pretrain's sarcasm was more direct: many job titles are just packaging for 'people who talk to LLMs all day.' Martin's take was more constructive: engineers are still needed, but engineers need to do engineering, not just write code. Aten Protocol singled out manufacturing's +13%: someone has to make things. Kevin Dorfer's judgment was that lower construction costs usually mean more things get built, not fewer builders. That judgment is optimistic, but it aligns with the direction of the charts. There were dissenting voices too. A senior developer said finding a job is still hard; one reply suspected the charts were propaganda; another mentioned Microsoft's layoffs.

Putting these voices together, the safer conclusion is that a rebound in total hiring and a cold individual job-search experience can both be true at the same time. The structure of jobs is changing, the entry points are narrowing, and skill requirements are being rewritten. a16z's charts can prove that demand for builders has not disappeared; they cannot prove that everyone can find a job more easily. What is worth watching next is the seniority breakdown, job reposting rates, and the salary distribution after AI skill requirements spread. Looking only at layoff announcements will miss the roles that were never posted.