Google Cloud Next, Google and Spotify Layoffs and Non-Compete Agreements | 🎙️#36
What are the consequences of Google fixating on Gemini? When is it ok to fire thousands of people? How stupid are non-competes? Join Pablo, Leo and Kirill for another discussion on “DevOps Accents”.
- Why there are not that many new features for GCP?
- What are Google goals with AI?
- What happens when you fire your Python team?
- Why is Spotify losing money?
- Should your janitor sign a non-compete?
- Is net neutrality a universally positive thing?
You can listen to episode 36 of DevOps Accents on Spotify, or right now:
In the realm of cloud computing and artificial intelligence, significant developments and concerns are shaping the future. As technology giants like Google invest heavily in new AI models, critical discussions are unfolding about the balance of innovation and traditional cloud infrastructure's integrity.
Google's Gemini Focus Raises Concerns
At Google Cloud Next 2024, the spotlight was on Gemini, Google's new AI model. Pablo expressed concerns about Google's intense focus on integrating Gemini into nearly all aspects of its cloud services. While AI enhancements can lead to advanced capabilities like improved security issue detection, there's a fear that this AI-centric approach might overshadow the foundational elements of cloud infrastructure—data storage, networking, and backend services.
I think the most important thing is that everything is AI. It doesn't matter whether it's a cloud company, an automotive company, or anything else; everything should be related to AI. The main point here, in this next era, is Gemini and the new model because they have a new model, and everything revolves around this new model. It's okay to have a new model, but the idea of a cloud environment and cloud infrastructure is the place where you will host your workloads, locate your data with your databases, and set up your network. At the end, it is the front end and the back end of your services, along with all these components, that define what a cloud is. Later on, you can have serverless components, you can have AWS Lambda, you can use Docker, or work with Kubernetes. There are many options, even reverting to the old way, known as on-prem, because it's still cloud-based computing. You can have your workloads there. However, the problem with AI these days is the vision of AI in the cloud; it's not just a new service. For sure, you have Vertex AI with Google, where you can use the models. But the idea is that they want to introduce generative AI as an enhancement to everything. So, it's cool to have Cloud Run, but Cloud Run is better with generative AI. In the end, generative AI is a large language model, and this has nothing to do with Cloud Run specifically, or with other components like Kubernetes. — Pablo Inigo Sanchez
I think maybe the idea from Google is that they will attract more customers, especially startups and upscale businesses, because of AI features. It's about the traditional sync with Google Cloud, right? People don't choose Google Cloud because it offers more services than AWS; they go to Google Cloud because it offers some services that are just way, way better. — Kirill Shirinkin
Kirill’s Perspective on AI Investments
Kirill highlighted the massive financial commitments by tech giants like Google and Meta towards developing AI technologies. While these investments are pivotal for technological advancement, there's a palpable tension between the costs and the tangible benefits these AI models offer. The competition in AI development could potentially lead to significant financial drain without guaranteed returns, especially for companies that might not emerge as leaders in AI.
But with these AI services, it's really hard to make money from AI. You have to invest so much money. And then, with these APIs like Vertex AI and Bedrock, I'm not sure what the margins are there, but you have to imagine it's way less than traditional compute reselling, or maybe some of them are even working in the opposite direction, just losing money. And that's what Meta said, you know, 'We've invested so much money, and we will invest many more billions, but there is not going to be any profit from this in the next years. We still need to invest more and more till we will be making money from this.' Which is okay for Meta because they have this big bet, but it's not okay for Google Cloud because they're saying, 'Okay, we cannot compete with AWS in terms of cloud infrastructure; we will now compete in the AI space.' But it's a very expensive game to play because AWS has way more money, way more income from selling cloud services than Google Cloud. So, they can afford to play this game a bit longer than Google. — Kirill Shirinkin
Leo Raises Alarm Over Google's Python Team Decision
Leo brought attention to Google's controversial decision to lay off its Python team as part of cost-cutting measures. This move could risk degrading the quality of Python development and maintenance, potentially impacting Google's AI and cloud services. The Python programming language plays a critical role in AI development, making this decision even more perplexing.
Anyone who's been to Munich knows it's not exactly a cheap city, but the savings from hiring cheaper labor might not outweigh the potential impact on the maintenance and development of Google's Python team. They essentially fired the whole team who had been working on that for years. These were the individuals who were literally on the Python Steering Council or long-time Python core developers. Replacing them with cheaper labor could lead to a loss of expertise and many more tools, as well as potential improvements that we've been discussing and anticipating. — Leo Suschev
Broad Implications of Spotify's Layoffs
The conversation also touched on the broader tech industry, with Spotify's recent layoffs serving as a case study of how reducing workforce can affect service quality and operational capabilities. The layoffs at Spotify, which were intended to cut costs, reportedly led to significant operational challenges and could hint at deeper issues within tech companies striving to balance profitability and service quality.
This is a trade-off, right? So if you want to be profitable, you have to lay off skilled, experienced technical experts—hundreds of them—and essentially compromise the quality of your own product. You end up making a more mediocre product, but you can stay afloat and be profitable. Now, this is the trade-off: if you want to make a good product that you can enjoy, where the algorithms work great, you have to pay a lot of money to good engineers, and if you do that, you are not profitable. So, it's a balance and it's understandable because, in the end, this is something that you, Pablo, like to mention a lot and tell me: 'This is a business, and they care about money.' I can understand that. Yet, in the end, they are making not the product that I liked in order to stay profitable. And I don't think this is the correct way, the right way to do the job. — Leo Suschev
The Debate Over Non-Compete Agreements
Non-compete agreements were another hot topic. Recent moves by regulatory bodies like the FTC to limit these agreements could significantly impact the tech landscape by enhancing job mobility for tech professionals. This shift could democratize innovation and reduce the monopolistic control companies have over their employees' careers.
Because an employer isn't your father, you know. It feels like they treat you, as a worker, like a stupid child. It's not okay. They say, 'I am your employer. I teach you everything. Now you cannot leave me.' It's almost like slavery because, in the end, if you are with me and you learn everything from me, then sorry, but you are mine. And if you want to leave, you can leave me, but you will never work for anyone else. — Pablo Inigo Sanchez
Net Neutrality: A Continuing Struggle
Finally, the discussion turned to net neutrality, a principle advocating for equal treatment of all data on the internet. With its reinstatement in the USA under the Biden administration, net neutrality is poised to ensure that internet service providers treat all data flows equally, without discriminating or charging differently by user, content, website, platform, or application.
These discussions illuminate the complex interplay between advancing technology and maintaining robust, equitable infrastructure and labor practices. As companies navigate these waters, the decisions they make today will likely resonate well into the future, impacting everything from AI development to employee rights and internet fairness.
Podcast editing: Mila Jones / milajonesproduction@gmail.com