Layoffs due to AI are coming - Weekly News Roundup - Issue #448
Plus: humanoid robots are getting to work; GPT Store launches next week; Intel launches a new AI company; bioengineered protein could enhance memory; Mamba and OnlyBots; and more!
Welcome to Weekly News Roundup Issue #448. As we step into 2024 (Happy New Year!), the main story of our first issue will focus on the not-so-happy prospect of more layoffs, this time due to AI replacing jobs traditionally done by humans.
In other news, humanoid robots are getting to work. Intel has launched a new AI company, and a new model named Mamba emerges as a potential competitor to Transformers. Additionally, AI bots have now got their very own social media platform, OnlyBots.
2023 was not a good year for tech workers. Almost every week, there was news about yet another company laying off a sizeable chunk of their workforce. At the beginning of the year, we saw tech giants like Amazon, Microsoft, Meta and Google laid off thousands of employees, setting a trend soon followed by other tech companies of various sizes, as documented by this timeline published by Computer World. According to Layoffs.fyi, a website tracking tech layoffs since the pandemic, in 2023 over 262,000 people in tech lost their jobs. That’s 100,000 more lost jobs than in 2022.
These mass layoffs coincided with the emergence of large language models and their dramatic entrance onto the public stage with ChatGPT and similar services. However, contrary to what some believed, these layoffs were not driven by the availability of AI tools (at least initially). The mass layoffs in 2023 were primarily due to tech companies adjusting after overhiring during the pandemic and responding to an economic downturn. In some cases, there was also pressure from investors to reduce the workforce to increase the value of the company’s stock. Many of these layoffs had been planned weeks, if not months, in advance, and thus did not consider recent developments in AI.
However, 2024 might paint a different picture. With over a year since ChatGPT's release, companies have had enough time to assess the impact of generative AI on their operations. Reports of firms replacing some employees with AI tools began surfacing mid-year. As a result, AI is increasingly becoming a factor in layoffs. According to a survey by Resume Builder, which asked 750 business leaders about their current or planned use of AI, 37% of the companies using AI said that they replaced workers with AI last year. Now, after experimenting with these tools, more companies are ready to pull the trigger on AI in 2024.
One of the first companies to replace part of their workforce with AI in 2024 is Duolingo. According to this post on Reddit, Duolingo has apparently laid off a “huge percentage of their contractors who did translations” in favour of using AI to do their job. Duolingo was a launch client for GPT-4 which powers the Duolingo Max service.
Another massive layoff due to AI might be at Google. Reports suggest that the tech giant is considering a reduction in its ad sales team, which puts 30,000 at risk. This decision follows the increased use of Performance Max, an AI-driven campaign planning tool launched in 2021. During the 2023 Google I/O event, the company announced enhancements to this ad tool, integrating generative AI capabilities that simplify the creation of custom assets. This development has led to a diminished need for specialized employees dedicated to ad sales in specific Google services, such as YouTube, Search, Display, Discover, Gmail, and Maps.
And this could be just the beginning. The same survey by Resume Builder mentioned earlier revealed that 4 in 10 companies said that AI is likely to replace workers in 2024 and 44% of companies surveyed say AI will lead to layoffs in 2024. For companies that plan to start using AI in 2024, the top anticipated uses include customer support (62%), creating summaries of meetings or documents (57%), and research (52%).
Last year, Goldman Sachs released a report estimating that 300 million jobs worldwide are at risk of being lost or diminished due to AI. According to Goldman, while automation may lead to job loss, it also spurs innovation, which in turn can lead to the creation of new job types. This innovation and automation are expected to result in cost savings for companies. By reallocating these saved resources towards business development and growth, the report suggests a potential increase in annual global GDP by 7%.
However, the short-term visibility of these gains remains uncertain, as highlighted in a report published in Harvard Business Review. The authors point out that AI is transforming workplaces at such a rapid pace that job losses are likely to precede the benefits, with white-collar workers being particularly vulnerable in the short term. The article proposes two approaches to mitigate this risk: one involves government intervention through welfare programs and retraining initiatives, and the other suggests a more proactive role for companies. This latter approach goes beyond mere task automation with AI, advocating for its use in enhancing employee productivity and generating new value, calling for a radical overhaul of corporate processes and a cultural shift towards agility and innovation. The report also emphasizes that for AI to be truly advantageous in the long term, companies must embrace a startup mentality and a bold approach, focusing on creating substantial value and new job opportunities. This strategy could help balance short-term job losses and avert long-term unemployment.
If you enjoy this post, please click the ❤️ button or share it.
I warmly welcome all new subscribers to the newsletter this week. I’m happy to have you here and I hope you’ll enjoy my work.
The best way to support the Humanity Redefined newsletter is by becoming a paid subscriber.
If you enjoy and find value in my writing, please hit the like button and share your thoughts in the comments. Additionally, please consider sharing this newsletter with others who might also find it valuable.
For those who prefer to make a one-off donation, you can 'buy me a coffee' via Ko-fi. Every coffee bought is a generous support towards the work put into this newsletter.
Your support, in any form, is deeply appreciated and goes a long way in keeping this newsletter alive and thriving.
🦾 More than a human
Brain stimulation could help doctors learn to use surgery robots
A study showed that applying gentle electric currents to the cerebellum enhances surgeons’ ability to transfer skills from virtual reality training in controlling surgery robots to real-world scenarios. Participants who received continuous brain stimulation were better at performing complex surgical tasks in both virtual and real settings compared to those with minimal or no stimulation. This research suggests that noninvasive brain stimulation could expedite skill acquisition in various fields reliant on virtual training, potentially reducing training time and resources.
🧠 Artificial Intelligence
▶️ Mamba: Linear-Time Sequence Modeling with Selective State Spaces (40:40)
Mamba is a new model that caught a lot of attention in the AI research world. It has the potential to be a competitor to Transformers, which are currently powering state-of-the-art large language models, as it scales better and can handle longer sequences more effectively. In this video, Yannic Kilcher thoroughly reviews the paper describing Mamba, explaining how it works and how it compares to other architectures such as Recurrent Neural Networks (RNN) or S4. For a more condensed explanation of how Mamba operates, check out this 16-minute video that also highlights the foundational papers upon which Mamba was built.
OpenAI’s app store for GPTs will launch next week
Yesterday, Thursday, January 4th, OpenAI sent an email to its customers announcing that the GPT Store will launch next week. Announced at OpenAI Dev Day in November 2023, GPT Store is a marketplace for GPTs - small and specialised versions of ChatGPT created by ChatGPT Plus users. The launch was initially scheduled for an earlier date but was postponed due to recent drama around OpenAI leadership.
Intel and DigitalBridge Launch Articul8, an Enterprise Generative AI Company
Intel is launching a new company, Articul8 AI, focusing on artificial intelligence software, with investment from DigitalBridge Group and others. The new company will be independent, with its own board, and Intel will remain a shareholder. Articul8 will focus on developing “a full-stack, vertically-optimized and secure generative artificial intelligence software platform”. This move is part of Intel's strategy to seek external funding for its business units, including plans for Mobileye Global and its programmable chip unit, according to Reuters.
OnlyBots: The world's first social network for bots only
OnlyBots is a social media network that implemented the Dead Internet theory (a theory suggesting the Internet is already overrun by bots). On OnlyBots, humans are not allowed. You can’t create an account or post anything. The platform is inhabited only by bots. You can create your own bot by describing in a sentence or two its personality from which the platform will generate the bot’s profile, and then you let it go to write its own posts, create images, and interact with other bots’ posts.
Images altered to trick machine vision can influence humans too
It has been known for some time that image recognition systems can be fooled into misclassifying images that look identical to us but were subtly altered. These “adversarial images” undergo minor pixel modifications, yet significantly impact AI classifications. Surprisingly, these small changes can also bias human decision-making, even when the alterations are nearly imperceptible, as researchers from Google DeepMind report. Although the results with humans are not as dramatic as with AI systems, which can be fooled to classify an image of a turtle as a banana, this discovery shows a similarity between human and machine vision. It highlights the need for further research to understand the influence of adversarial images on both AI systems and human perception.
Does China Want Generative AI?
In this article at
Making an image with generative AI uses as much energy as charging your phone
A study by Hugging Face and Carnegie Mellon University reveals that generating an image with AI can consume as much energy as fully charging a smartphone, while text generation is less energy-intensive. The research highlights that AI's carbon footprint is largely due to its usage, not just its training. It suggests a need for more specialized, less energy-consuming AI models and raises awareness about the carbon emissions associated with everyday AI usage.
If you're enjoying the insights and perspectives shared in the Humanity Redefined newsletter, why not spread the word?
Humanoid Robots Are Getting to Work
2024 could be a big year for humanoid robots. A handful of well-funded companies will be putting their robots through pilot programmes to assess their effectiveness in real-world scenarios. One of such robots, Agility Robotics’ Digit, is already being tested by Amazon. This article from IEEE Spectrum provides an insight into these companies, their offerings, and the challenges they face. For a deeper dive into the companies at the forefront of commercial humanoid robotics, check out my article Ten Companies Leading the Upcoming Humanoid Robot Wave.
Fourier Intelligence launches production version of GR-1 humanoid robot
Fourier Intelligence recently announced that it is ready to launch the production version of the GR-1 humanoid robot. The Singapore-based company plans to market the GR-1 for various applications, including research, education, healthcare, and household services, leveraging its established sales channels from its rehabilitation offerings. No price has yet been announced for the production unit.
The war in Ukraine and the conflict between Hamas and Israel have demonstrated how cheap drones have revolutionized the battlefield. To better prepare for drone-filled battlefields, the US military has opened the Joint Counter-small Unmanned Aerial Systems University. The academy will train up to 1000 soldiers annually in dealing with enemy drones, employing methods ranging from shooting them down with traditional rifles to employing jamming devices.
Canada approves Pfizer's gene therapy for bleeding disorder
The Canadian health regulator has approved Pfizer's gene therapy for treating haemophilia B, a rare inherited bleeding disorder. This approval is based on late-stage trials demonstrating the therapy's superiority over the current standard of care, which involves replacing a blood-clotting protein known as factor IX. The therapy, branded as Beqvez, comes after the U.S. Food and Drug Administration's approval of CSL's Hemgenix in November 2022, the first one-time gene therapy for haemophilia B.
Bioengineered protein could enhance memory
Neurologists have identified a direct link between memory and the activity of a brain protein, LIMK1. By genetically modifying LIMK1 in mice, researchers improved memory formation and slowed cognitive decline. This discovery shows that altering LIMK1's structure enhances synaptic information transmission, and opens potential pathways for treating memory-related conditions like Alzheimer's and dementia.
Cyborg computer combining AI and human brain cells really works
Researchers at the University of Indiana Bloomington developed "Brainoware," a biohybrid computer combining a brain organoid and traditional AI. This system demonstrated a 78% accuracy in speech recognition tasks using neural responses from the organoid. While less accurate than conventional AI systems and requiring significant power resources, this proof-of-concept highlights the potential for human biology to enhance computing capabilities.
Thanks for reading. If you enjoyed this post, please click the ❤️ button or share it.
Humanity Redefined sheds light on the bleeding edge of technology and how advancements in AI, robotics, and biotech can usher in abundance, expand humanity's horizons, and redefine what it means to be human.
Humanity Redefined is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
A big thank you to my paid subscribers, to my Patrons: whmr, Florian, dux, Eric, Preppikoma and Andrew, and to everyone who supports my work on Ko-Fi. Thank you for the support!