Useful information
Prime News delivers timely, accurate news and insights on global events, politics, business, and technology
Do you want smarter ideas in your entrance tray? Register in our weekly newsletters to obtain only what matters to the leaders of AI, data and business security. Subscribe now
Software developers spend most of their time No Writing code; The recent industry research found that real coding represents as little as 16% of developer’s work hourswith the rest consumed by operational and support tasks. As engineering equipment is pressed to “do more with less” and CEO boast of how much of their base is written by AI, a question remains: what is done to optimize the remaining 84% of the tasks in which engineers are working?
Keep developers where they are the most productive
A great culprit for developer’s productivity is the change of context: the constant jump between the growing range of tools and platforms necessary to build and send software. A Harvard commercial review study found that the average digital worker changes between applications and websites almost 1,200 times a day. And each interruption matters. California University discovered that it is needed about 23 minutes recover the focus after a single interruption completely and, sometimes, worse, like almost 30% of Interrupted tasks never resume. The context switch is located in the Dora center, one of the most popular performance software development frames.
In an era in which companies promoted by AIs are trying to empower their employees to do more with less, beyond “only” giving access to large language models (LLM), some trends are emerging. For example, Jarod Ruhland, main engineer of Brex, hypothesis that “developers deliver their greater value when they focus within their integrated development environment (IDE)”. With that in mind, he decided to find new ways of making this happen, and the new Anthrope protocol could be one of the keys.
MCP: A protocol to bring context to IDES
Coding attendees, such as IDES with LLM food such as cursor, co -pilot and Windsurf, are in the center of a Renaissance of the developer. Its adoption speed is invisible. Cursor became the fastest growth in history, reaching $ 100 million RU within 12 months after the launch, and 70% of Fortune 500 Companies use Microsoft Copilot.
Ai scaling reaches its limits
The power limits, the increase in token costs and inference delays are remodeling Enterprise AI. Join our exclusive room to discover how the best teams are:
- Convert energy into a strategic advantage
- Efficient inference architecture for real performance profits
- Unlock competitive roi with sustainable AI systems
Ensure your place to stay at the forefront: https://bit.ly/4mwgngo
But these coding attendees were only limited to the context of the code base, which could help developers write code faster, but could not help with the change of context. A new protocol is addressing this problem: model context protocol (MCP). Run in November 2024 by Anthrope, it is an open standard developed to facilitate integration between artificial intelligence systems, particularly LLM -based tools and external data tools and sources. The protocol is so popular that there has been a 500% increase of new MCP servers in the last 6 months, with an estimate of 7 million downloads in June,
One of MCP’s most shocking applications is its ability to connect the coding assistants directly to the tools in which developers trust every day, rationalizing workflows and drastically reducing the change of context.
Take the development of characteristics as an example. Traditionally, it implies bouncing among several systems: reading the ticket on a project tracker, looking at a conversation with a teammate to clarify, find the documentation of the details of the API and, finally, open the ID to start encoding. Each step lives in a different tab, which requires mental changes that slow down developers.
With MCP and modern assistants such as Claude de Anthrope, all this process can occur within the editor.
For example, the implementation of a function, all within a coding assistant, becomes:
The same principle can be applied to many other workflow engineers, for example, an incident response for sres might seem:
Nothing new under the sun
We have seen this pattern before. During the last decade, Slack has transformed productivity in the workplace by becoming a center for hundreds of applications, which allows employees to administer a wide range of tasks without leaving the chat window. The Slack platform reduced the change in context in everyday workflows.
Riot Games, for example, connected around 1,000 Slack applications, and engineers saw a 27% reduction With the necessary time to prove and iterate the code, a time of 22% faster to identify new errors and a 24% increase in the features launch rate; All were attributed to rationalizing workflows and reducing friction of tools.
Now, there is a similar transformation in software development, with AI attendees and its MCP integrations that serve as a bridge to all these external tools. Indeed, the IDE could become the new command center all in one for engineers, just like Slack has been for general knowledge workers.
MCP may not be ready for the company
MCP is a relatively nascent standard, for example, Security Wisem MCP does not have an incorporated authentication or a permit model, which depends on external implementations that continue to evolve, there is also ambiguity on identity and audit: the protocol does not clearly distinguish if a user or the AI itself caused an action, which makes responsibility and the control of access without personalized solutions difficult additional Lori Macvittie, distinguished engineer and main evangelist at the F5 Networks office of the CTO, says that MCP is “breaking central security assumptions that we have had for a long time.”
Another practical limitation arises when too many MCP tools or servers are used simultaneously, for example, within a coding assistant. Each MCP server announces a list of tools, with descriptions and parameters, which the AI model must consider. The floods of the model with dozens of available tools can overwhelm your context window. The performance is degraded significantly as the tool count grows with some IDE integrations has imposed hard limits (about 40 tools in the IDE cursor, or 20 tools for the OpenAi agent) to prevent the application from swelling beyond what the model can handle
Finally, there is no sophisticated form for the tools to be automatically discovered or suggested contextually beyond enumerating them all, so developers often have to alternate manually or cure what tools are active to keep things working without problems. Referring to that example of riot games that install 1,000 Slack applications, we can see how it might not be suitable for business use.
Less turn serve, more software
The last decade has taught us the value of bringing work to the worker, from slack channels that address the updates to the email methodologies of “Zero Tray” and the engineering panels of unified platforms. Now, with ia in our tool kit, we have the opportunity to train developers to be more productive. Suppose Slack became the Business Communication Center.
In that case, coding assistants are well positioned to become the center of software creation, not only where the code is written, but where all contexts and collaborators join. By keeping developers in their flow, we eliminate the constant change of mental equipment that has affected engineering productivity.
For any organization that depends on software delivery, take a look at how its developers spend their day; You may be surprised what you find.
Sylvain Kalache leads ai labs in Rooted.
