Welcome to Runtime! Today: Stack Overflow's annual developer survey helps sort enterprise tech hype versus reality, AMD takes aim at a new rival, and the latest funding rounds raised by enterprise tech startups.
Their stacks overfloweth
All of us office drones have endured technology tools we are forced to use at work that don't exactly spark joy, but are grudgingly accepted because they get the job done. But the world is changing, and employees are exerting more control over the tech they use on the job.
Developers and other tech professionals enjoy a very competitive market for their services, and tend to have strong opinions about enterprise tech tools. Stack Overflow's annual survey of those people is a must-read to get a handle on what technologies they're actually using, and which tools they'd rather be using.
The top technologies in use among those surveyed last month — 89,184 tech professionals, mostly in the U.S. — haven't changed significantly over the last several years.
- But PostgreSQL took over the most widely used database crown, edging out MySQL, although both are in use by more than 40% of survey respondents.
- Visual Studio Code was the single most-used product on the list, serving as the integrated development environment for 74% of respondents.
- Nearly 72% of professional developers said their companies were using CI/CD tools, the largest category on the list. Automated testing and DevOps tools are also widely used.
But when tech pros were asked which tools they actually want to use more frequently after getting a taste, some interesting results appeared.
- Rust users were more enthusiastic about writing future code in that programming language than any other language that appeared on the list; interest in Go, a similar memory-safe language, wasn't too far behind.
- PostgreSQL looks set to maintain its status as the most widely used database, with more people saying they want to continue using it in the future among all listed databases, but Datomic and DuckDB's small user bases are also extremely enthusiastic about those products.
- And Vercel is also winning converts, a sign that the platform modernization movement is probably here to stay.
Questions about AI technologies were placed in a separate category this year, which makes sense given the immense amount of discussion that has surrounded the generative AI boom.
- Among professional developers, 70% said they will be using AI software development tools this year, and 44% are already using them.
- But they're not sure how much they can trust the output of those tools; 39% said they "somewhat trust" the tools, while 22% said they "somewhat distrust" them and 30% weren't sure what to think.
- Yet almost 83% of those currently using AI tools are using them to write code.
The entire survey is worth a look for deeper insights into how enterprise tech tools are being used in the field. One of the things that makes enterprise tech interesting is that there is no One Way To Build Software, so technologies that fly a little below the radar can still be incredibly important and valuable.
- Look for profiles of the companies, developers, and users behind some of those lesser-known but coveted technologies in future editions of Runtime.
AMD looks up at a new target
AMD's re-emergence as a chip power player after years of dysfunction arrived on the back of its server processors, which took share from Intel thanks to their performance and Intel's own dysfunction. But a different chip market is shaping up to be the engine of the '20s.
AMD shed a little more light on its MI300 processor family, which will begin shipping later this year, at an event in San Francisco Tuesday. "None of AMD’s competitors have (or will have) a combined CPU+GPU product like the MI300 series this year," according to Anandtech.
Nvidia introduced a hybrid GPU/CPU design several years ago, citing the growing importance of AI-related tasks handled by GPUs to the overall performance of the data center, which has been traditionally ruled by the CPU. But that design used two chips connected by an interface, whereas AMD's MI300 is a single chip that combines both types of processing horsepower.
"MI300 will not quite be a make-or-break product for the company, but besides getting the technical advantage of being the first to ship a single-chip server APU (accelerated processing unit) … it will also give them a fresh product to sell into a market that is buying up all the hardware it can get," Anandtech said. Nvidia is currently enjoying most of the benefits of that market frenzy and will join AMD with a single-chip server design later this year, while Intel just canceled its take on that design and looks further and further behind its data-center competitors.
Blackpoint Cyber landed $190 million in Series C funding to expand the managed security services it provides to companies that manage tech infrastructure for other businesses.
Cohere raised $270 million from Nvidia, Oracle, Salesforce, and existing investors to compete with OpenAI in developing AI foundation models.
Mistral AI, another would-be OpenAI competitor based in Europe, raised a €105 million seed round (!) and promptly launched the company.
Digibee raised $60 million to support its low-code approach to helping customers build integrations across their applications.
The Runtime roundup
That much-ballyhooed decision by Amazon Prime Video to move from a microservices architecture to a monolithic design was actually the creation of a more sophisticated approach to microservices, according to The New Stack.
Thanks for reading — see you Thursday!