Lazy Sunday #7

Good people! Sunday!

I am still a little bit star-struck today. Yesterday I went to see Dr. Andrew Hubermann at the Sydney Opera House.

His ability to explain complex scientific facts in an approachable manner, mixing in a good portion of humour while going from stress to focus to cuttle fish (ok, he diverted there a little bit), was absolutely great.

I will share my notes on Thursday, need to decipher the multiple pages of scribbles first.

Find my research

Social cost of AI performance management

Algorithms and machine learning are continuously leveraged in HR departments, from hiring to performance management. A study in Germany found, that algorithmic performance management reduces social interactions between co-workers. In other words, if employees know that their performance and appraisal is in done in tandem with or by an algorithm based on data, employees collaborate, interact and assist each other less. Employees feel objectified.

As algorithmic performance evaluation will probably not disappear, the researchers argue companies can offer more settings for social interactions in the offices and warehouses. Which would follow a trend of the office becoming an interaction and collaboration space and less a “sit-alone-in-front-of-screen”-space. (HBR)

Pillars of strategic agility

This study I recommend every founder and manager to read (yourself or your favourite LLM model, your choice).

The underlying question: what makes an organization agile in today’s competitive and workforce dispersed world?

Three pillars, I’d highlight:

1. Knowledge Management: this has been a challenge in every organization I worked with or for. While collecting information through wikis or others seemed doable for most, providing them in a logical structure that allows employees to FIND the information, has been a totally different beast. No employee will look for information for more than 10 minutes, let alone the impact it has on productivity. But, with more powerful LLM’s we might enter an era, where knowledge within companies can be more easily accessed than ever.

2. Talent Management: learning and career advancement is on everyone's agenda and most often executed poorly. I think we need new tools that allow for bite-sized context relevant learning. If 1) can be mastered, learning and development becomes much easier. Also, make sure you cut down the silos. People need to be able to look over each other's shoulders to learn and get back with new ideas.

3. Innovation: this is not just about creativity but also about the ability to engage the right partners and knowledge sources to create unique value. “Do-it-yourself” is over. Co-development and co-distribution will become a core model for high performing companies.

Additionally, the study mentions digitization and sustainability. For the former, I say: duh, really? For the latter, I say: sure, ok! It is a must in today's environment, but I am not fully convinced it provides agility, but it is important for institutional investors.

Definitely give it a read! (Wiley)

Got a kiddo? Learning to read? Here are some tips!

I might be heavily biased, as my son is in the process of learning how to read. As we are a bilingual family, it is hard to stop him trying to read both, and getting frustrated in the process. Him, not me.

The following article is a great lesson in learning in general: focus on the dots first, then connect it to lines to see the larger picture. The process outlined for reading, can really be applied to other situations as well. I found it worth a read, I think, even if you don’t have to teach reading yet. With a little bit of abstraction, you will find some applications for yourself. I am sure. (ABC)

On top of tech

Adidas gets neuroscientific

Adidas works with neuro11 to identify ways to help new athletes deal with stress better. I am telling you, neuroscience like this is getting into the mainstream and opens real, new possibilities to improve performance, getting into a state of mind before difficult situations and relaxation before / after games. (Adidas)

3D-printed brain tissues

The first brain tissue, that can grow and function like a typical brain tissue, has been 3D printed at the University Of Wisconsin. 🤯 (ScienceDaily)

GPT3-enabled chatbot as therapist for students

Suicide is one of the most common reasons for death in the 15-29 age bracket. Insecurities about one's self, about social belonging, there are many reasons.

Further, it is very challenging to open up, talk about the feelings and look for help from other people. There is an abstraction of engagement with bots. Even though everybody knows it is a machine, people open up and engage such models earlier and more transparently.

Enter Replika, a GPT-3 based chat, that can be engaged in difficult situations. The conversational-like style GPT models allow nowadays, can get much more facetted both in conversation style and responsiveness to nuances.

Replika is being used by more than 25 million students in the states, and, based on user feedback, helped many to not take their lives.

Immense impact! (Nature)

Sleepy dollars are on the rise

Sleep is probably one of the most important aspects of brain health and thus performance. StellarSleep raises a significant $6 million in seed (not a bad round) to provide a “sleep psychology” to help with chronic insomnia. While the $60 monthly user fee seems pretty high in comparison to other subscriptions like Oura’s $5, the alternative would be to try and find a sleep clinician - which can take many months. (Techcrunch)

The sheer amount of sleep tech is immense. Last year Samsung Ventures invested in Earable, a company that developed a sleep wearable to help fall and stay asleep.

Onera has raised $32 million earlier this month to provide tech for sleep studies at home. If you have insomnia or sleep apnoea (snoring), you will have to find a clinic, sleep a night or two there, while being hooked up to cables, and then discuss the data and next steps with your doctor. Many clinicians, as well as patients, report their sleep to be worse during sleep because you are not in your bed, not surrounded by your environment, but hooked up to cables. Onera brings the tech home. (NextDigitalHealth)

EarlyBird Ventures raised a $173 million EUR fund for invest in patient-centric tech

Earlybird, founded in 1997, is one of the leading venture capital firms in Europe. In 2016, Earlybird raised an €85 million health fund, one of its first in Europe. Now, not even 10 years later, Earlybird raises Fund II with €173 million, though targeting slightly later stage companies. It will be interesting to see where this dry powder will be used. (FierceBioTech / EarlyBird LI)

Media Box

Work and workplace trends in 2024

Regularly, the World Economic Forum gathers data about all kinds of trends and risks. In the last “Global Risk Report 2024” and “Chief Economist Outlook 2024”, WEF explores the workplace (amongst others).

1. AI will increase productivity, but not everywhere

2. Digital knowhow is in demand

3. Unemployment will continue to be volatile

4. Office spaces will become more pop-up on demand

5. AI will need to be taught and new acquiring new skills even more important

6. More women at work, but gender gaps (e.g. pay) persist

Can we compare LLM’s and the brain?

I am not going to repeat that 2023 was a bonker year for AI and LLM’s (oops, I did it), and each additional new tool, application and model, adds fuel to the burning fire of philosophers to understand how far away are LLM’s in their current form from our own mind and brain.

As our brain can anticipate and predict, internalize feedback, learn and get better, how much of a different, and more evolved, process is that compared to the “simplicity” of LLM’s? Have LLM’s already caught up, with the occasional hallucination to distract from their real capabilities? I am not for or against any of these arguments really, but it shows how little we do actually know about the brains’ features, if we find ourselves pondering over the proximity of technology to our own minds?! (TIME)

Misc but not least…

If you have a Twitter account (sorry, X account), please have a look and compare the Sora and Midjourney. Remarkably similar. Is this a function of the similarities of the algorithms or the training content? (Link)

That was quick!

Sunday, done. Learnt something new!

Ready for the week ahead.

Awesome!

Alex