Obviously, the world of work for technologists is in flux. This year’s data capture a workforce navigating serious uncertainty. Layoffs have become a perennial industry headline, AI is exploding expectations of what a single person can get done, and the traditional employer-employee relationship is evolving in response to these seismic shifts.
Many-to-one: can we do more by ourselves than with a team?
More respondents in this year’s survey are striking out on their own. The percentage of respondents working as a company of one—that is, freelancers and sole proprietors—more than doubled this year from 4% to 10%. That’s a huge leap for one year, and it’s not the kind of jump you can chalk up to gradual market forces. Two explanations, by no means mutually exclusive, come to mind. One is opportunity: AI tools have made it possible for many types of tasks to be done by a single worker. OpenAI research finds among all workspaces’ non-generic messages, 43.5% fall outside of the user’s occupation. Snowflake’s Head of Developer Experience Umesh Unnikrishnan also pointed out how job duties have compressed. “I started my career at Microsoft. I was responsible for the folders feature in Outlook. But now someone like that could build something like Microsoft Outlook by themselves.”
10% work for themselves
The percentage of people saying it was just them in an org jumped from 4% to 10% this year. AI has made it easier for people to do work outside of their primary expertise, so it could point to the rise of solopreneurs, full-stack freelancers, or devs finding alternate work arrangements after layoffs.
View and share | Respondents | Percent | |
|---|---|---|
| 10,000 or more employees | 2,941 | 13.9% |
| 5,000 to 9,999 employees | 1,017 | 4.8% |
| 1,000 to 4,999 employees | 2,409 | 11.4% |
| 500 to 999 employees | 1,307 | 6.2% |
| 100 to 499 employees | 3,474 | 16.5% |
| 20 to 99 employees | 3,677 | 17.4% |
| Less than 20 employees | 3,555 | 16.8% |
| Just me - I am a freelancer, sole proprietor, etc. | 2,213 | 10.5% |
| I don’t know | 522 | 2.5% |
New tools, same problem
That tension—opportunity versus necessity—shows up again when we look at how respondents feel about the jobs they do have. Satisfaction is down: more developers report being unhappy at work than last year (up 4pp), and fewer report being genuinely happy (down 2pp). The job market has cooled, a New York Times investigation points to a new phenomenon that could be exacerbating the job satisfaction trends in the US at least and that is the “quality jobs crisis”. Satisfaction data doesn't tell us why any individual person feels the way they do, of course, but the survey did ask what aspects of work have contributed to a person’s job feeling easier or harder to do. The main take-away is work remains complex and messy: the most selected reasons jobs got harder were “unclear priorities”(17%) and “inefficient processes”(15%). Tellingly, 21% of respondents say nothing at all has made their job easier in the past six months.
Unclear priorities or requirements are the reason 17% say their job has gotten more difficult
Ambiguous direction serves as the primary operational friction point, closely followed by inefficient workflows (15%) and time/staffing shortages (14%).
View and share | Respondents | Percent | |
|---|---|---|
| Unclear priorities or requirements | 3,029 | 16.6% |
| Inefficient processes, workflows, or coordination | 2,658 | 14.6% |
| Not enough time, staffing, or resources | 2,566 | 14.1% |
| Leadership activity or organizational changes | 2,162 | 11.9% |
| Nothing has made my job harder in the past 6 months | 1,743 | 9.6% |
| Unreliable systems, infrastructure, or environments | 1,523 | 8.4% |
| Poor documentation or knowledge sharing | 1,292 | 7.1% |
| Other | 1,108 | 6.1% |
Tools are the largest positive intervention, yet slightly more (21% vs. 20%) say nothing made work easier in the past year
View and share- Nothing made it easier
- The right tools
- Supportive teammates
- Clear priorities
- Better processes
| Respondents | Percent | |
|---|---|---|
| Nothing has made my job easier in the past 6 months | 3,902 | 21.1% |
| Access to the right tools or platforms | 3,714 | 20.1% |
| Supportive teammates, managers, or mentors | 3,191 | 17.2% |
| Clear priorities or requirements | 1,314 | 7.1% |
| Better processes, workflows, or coordination | 1,256 | 6.8% |
The picture painted by the data is that technologists are being forced to adapt on two fronts at once.
The shape of employment itself is allowing for more independent, solo work, while complexity and uncertainty are making teams grow more cautious and discerning about what actually helps them at work. Through it all, the data suggests, modern employees working in technology have grounded expectations for what AI can do for them, even as the tech slowly wins their trust.
Users' top reasons to endorse a technology are reliability and performance
Core reliability and performance are the leading reasons professionals endorse a technology, taking priority over pure productivity gains (16%) and ecosystem strength (13%).
View and share | Respondents | Percent | |
|---|---|---|
| It is reliable, performant, and stable | 3,874 | 17.3% |
| It helps me be more productive or save time | 3,563 | 15.9% |
| It has a strong community, ecosystem, or open-source presence | 2,888 | 12.9% |
| It has a strong developer experience, such as APIs, SDKs, tooling, or integrations | 2,541 | 11.4% |
| It is flexible, customizable, or easy to adapt to my needs | 1,678 | 7.5% |
| It has a reputation for quality and long-term trustworthiness | 1,666 | 7.5% |
| It has excellent documentation, examples, or learning resources | 1,370 | 6.1% |
| It is easy to learn and use | 1,226 | 5.5% |
| I’m not sure | 1,223 | 5.5% |
| Other, please specify | 632 | 2.8% |
| It has useful AI features or agent capabilities | 505 | 2.3% |
| It has strong security, privacy, or compliance features | 423 | 1.9% |
| It has manageable cost, pricing, or licensing | 421 | 1.9% |
| I would not recommend any technology I use | 361 | 1.6% |
Explore Work data
See the full breakdown