Key Takeaways
- The layoffs in big tech have had a dampening effect on psychological safety across the whole industry
- Generative AI tools like Co-pilot or ChatGPT can have a significant impact on developer productivity but are also significantly flawed
- Responsible tech is more than just regulatory compliance and companies are having to be more socially responsible in order to attract and retain customers and employees
- Asynchronous work is becoming much more accepted and organisations adopting the practices are gaining real benefits
- Hybrid workplaces are now the norm, it is very important to be deliberate about the reasons for bringing people together in person, rather than just having a fixed schedule
The two biggest factors influencing organisational culture in early 2023 are the highly publicised and hugely disruptive layoffs across the tech industry and the emergence of Large Language Models such as ChatGPT.
Layoffs in tech have damaged psychological safety
The biggest impact of the layoffs has not been the number of people out of work, but the widely publicised reports of the inhumane approach many of the large employers took to announcing and communicating the layoffs. Stories of people finding out they no longer have a job when they are not able to log on to their corporate account, or getting a text message from an unidentified number telling them to drop off their company equipment with no other communication have a very negative impact on the industry as a whole and psychological safety is down in almost every high tech company.
There is a realistic argument that the layoffs were a sensible response to the changing economic climate and that for many companies it is about right-sizing rather than down-sizing, however the way they were communicated to the impacted staff and to the world at large has resulted in a loss of trust and reduction of employee engagement across the industry. Rebuilding a generative culture will be difficult, but can be done with care and deliberate design.
Despite the large numbers who were laid off, there are still large shortages of tech workers around the world and most of those impacted are finding new roles relatively quickly.
Rebecca Parsons, CTO of Thoughtworks commented on the podcast accompanying this trend report that:
There is a strong business case for having a strong employee value proposition, because turnover is expensive and replacing people is expensive. If you treat people well, you can make all of these wonderful business case arguments, but you can also look at it as it's just the right thing to do.
Employers who treat people badly in layoffs are going to struggle to attract people when the economy turns and they look to grow again.
Large language models - friend or foe for developers?
The launch of large language models such as ChatGPT, the premium service ChatGPT Plus and the enhanced GitHub Copilot have resulted in headlines about jobs that are at risk of being replaced by AI, with programmers at the top of some lists. The question of how real the threat is, and how do the potentially displaced adapt is very open at the moment as it is early days in the evolution of the tools. Legal battles are shaping up around open source licensing and copyright both of the source material used to generate content and the copyright status of generated content.
On the other hand there are reports of how using these tools can significantly increase developer productivity - handling many of the more mundane tasks of coding, reducing the cognitive load and allowing the creative human to apply their energies to the areas of development where creativity is best applied. Developer tools already include AI assistance, and this will continue to increase.
What has become clear is that there is a skill to crafting the questions to ask; prompt engineering is the name most used to describe this skill. These code assist tools work best if you know the answer that you want. Users need to verify the results and validate that the responses actually address the issue raised. The models lack context and don’t know what they don’t know and that makes using them potentially risky.
ChatGPT recently had a temporary shutdown of its service due a bug which enabled users to view one another’s prompts. This illustrates the inherent risks of sharing confidential data with the providers of large language models. OpenAI’s terms of service clearly state that they collect information which is “included in the input” or in “file uploads.” As effective prompt engineering relies on providing contextual information, it is important that organisations safeguard themselves against inadvertent data privacy breaches. While some organisations have been putting in place governance, JPMorgan recently blocked all staff from submitting data to ChatGPT.
Unskilled cybercriminals are already using ChatGPT to generate malware; as technologists we need to be aware of the potential impacts and actively mitigate the risks.
While these two factors have dominated the headlines there has been a lot happening in other areas that is impacting organisational cultures and people operations.
Movements in responsible tech
Technologists are becoming more aware of the broader societal impacts of the work they do and the importance of ethical considerations in the products we build and the way we build them. Climate impact, social good, privacy, safety, diversity, equity and inclusion and the impact of bias need to be considered in our development processes.
Responsible technology goes beyond compliance with rules and regulations. Taking a responsible view means considering the potential impacts of the products we’re building before we start building, asking the hard questions about how could this product be applied in ways we don’t intend, who are the communities and stakeholders that we haven’t consulted or considered, how can we actively prevent harm, rather than just complying with the minimum regulatory requirements?
It has a financial benefit from enhanced reputation, from employees and customers wanting to be associated with ethical companies and from customers being prepared to pay more for ethical products. Responsible technology doesn’t just happen - it needs deliberate practice and clear goals covering areas such as supply chain, operations, employees and clients, in addition to the more obvious issues around product and service development.
DevSusOps is more than just being carbon neutral - to be effectively sustainable an organisation needs to look to the needs of all stakeholders, social, environmental and economic.
Ethical standards and guidelines for software engineering are evolving and emerging in new areas, such as agile coaching.
Project to product gaining traction
The ideas around project to product/#NoProjects are gaining traction as more organisations shift from seeing their IT departments as an overhead cost to truly partnering as a value adding service essential to running the business effectively and meeting customers’ needs.
Value stream management is being used more frequently to visualise and optimise the development process in end to end DevOps implementations.
As organisations adopt these approaches, project based working breaks down and product management becomes necessary and effective. Approaches such as Evidence Based Management provide tools to help maximise value delivered and focus on business outcomes.
The worth of an architect
The frequent conflict between development teams and architects has caused harm and some organisations have found ways to effectively reconcile the potentially competing goals. Development teams are tasked with delivering features as quickly as possible to add value and delight customers. Architects have a longer term view of maximising the long-term asset value of the IT estate. Aligning these two perspectives requires a shift in communication and engagement.
Making architecture a core part of the development process requires constant collaboration and communication, not through presentations and ivory-tower decrees, but through working code and practical application that show the value the big picture thinking has on the product both in the short and long term.
Organisation are increasingly distributing the architectural function across teams, through the use of cross-team reviewed Design Docs or RFCs. Grygoriy Gonchar recently published an article on InfoQ about the use of a Simple Framework for Architectural Decisions, which includes the use of RFC like Architectural Decision Records (ADRs). Gonchar wrote that such approaches ensure alignment with “business goals” and allow teams to make “informed” architectural decisions.
Team topologies for rapidly changing environments
Team topologies are emerging that support rapidly evolving organisations and the remote/hybrid workforce. Reteaming is a competency that organisations need to build - enabling teams to rapidly form around new challenges or opportunities and become fully productive and well gelled quickly.
Understanding the impact of Conway’s Law, and using the Reverse Conway approach enable organisations to deliberately align communication to be more effective. Team Topologies has provided patterns to guide organisations in restructuring teams to optimise for value delivery and minimising cognitive load. As Twitter’s mass layoffs have recently shown, the downside of mass redundancies can be the loss of moral, sustainability and stability due to increases in cognitive load for remaining teams.
The pace of change and the dynamic nature of employment mean that organisations need to continue reviewing the changing cognitive load on teams, when faced with large scale organisational change. Teams must also become good at onboarding new people to help them become productive.
Staff Plus and technical career pathways
The Staff Plus role is gaining more attention and a clear recognition that there are career paths which don’t require going into people management positions. Deliberately shifting into and out of people leadership roles on a pendulum is a viable option that both deepens and broadens skillsets.
Hybrid work will continue to evolve
The shift from fully remote to hybrid working continues steadily, with a few companies mandating complete back to the office and losing people as a result. Hand in hand with the shift in working comes a far greater acceptance of asynchronous work in remote teams and an uptick in the need for more effective written communication and the need for process and guidelines on how to (for instance) communicate architectural decisions so they are understood across the whole community who need to know about them. Finding the right balance between freedom & flexibility and documentation & process is an important factor when much of the work is done asynchronously.
Workplace design is changing as organisations retool for hybrid working. We see designs with fewer desks and more collaboration spaces with technology to support teamwork, pods and quiet places for individual work and quiet conversations and reflection.
Wellbeing and helping support mental wellness need to be taken into account when establishing ways of working. Understanding the diverse needs of the organisational community and ensuring inclusiveness and respect enables more effective collaboration.
The collaboration tools available today have improved significantly, but they still have gaps and frustrations. A hybrid meeting with some in-person and some remote participants is often an exercise in frustration with poor sound quality and missed interactions, exacerbating the us-and-them divide which can easily emerge in distributed teams.
Listen to the Trends Report Discussion on the InfoQ Engineering Culture Podcast
The Culture & Methods editorial team met remotely to discuss these trends and we recorded our discussion as a podcast. You can listen to the discussion and get a feel for the thinking behind these trends.