Google sets the bar for AI language models with PaLM. The number of parameters is important in LLMs, although more parameters don’t necessarily translate to a better-performing model. PaLM 540B is in the same league as some of the largest LLMs available regarding the number of parameters: OpenAI’s GPT-3 with 175 billion, DeepMind’s Gopher and Chinchilla with 280 billion and 70 billion, Google’s own GLaM and LaMDA with 1.2 trillion and 137 billion, and Microsoft – Nvidia’s Megatron–Turing NLG with 530 billion.

Google sets the bar for AI language models with PaLM. The number of parameters is important in LLMs, although more parameters don’t necessarily translate to a better-performing model. PaLM 540B is in the same league as some of the largest LLMs available regarding the number of parameters: OpenAI’s GPT-3 with 175 billion, DeepMind’s Gopher and Chinchilla with 280 billion and 70 billion, Google’s own GLaM and LaMDA with 1.2 trillion and 137 billion, and Microsoft – Nvidia’s Megatron–Turing NLG with 530 billion.

The first thing to consider when discussing LLMs, like any other AI model, is the efficiency of the training process. PaLM uses a standard Transformer model architecture, with some customizations. Transformer is the architecture used by all LLMs, and although PaLM deviates from it in some ways, what is arguably more important is the focus of the training dataset used. The dataset used to train PaLM is a mixture of filtered multilingual web pages (27%), English books (13%), multilingual Wikipedia articles (4%), English news articles (1%), GitHub source code (5%), and multilingual social media conversations (50%). This dataset is based on those used to train LaMDA and GLaM. Nearly 78% of all sources are English, with German and French sources at 3.5% and 3.2% and all other sources trailing far behind.

PaLM 540B surpassed few-shot performance of prior LLMs on 28 of 29 tasks. PaLM outperforms the prior top score of 55% achieved by fine-tuning GPT-3 with a training set of 7,500 problems and combining it with an external calculator and verifier. This new score also approaches the 60% average of problems solved by 9- to 12-year-olds — the target audience for the question set.

Now, let's shift gears and talk about the science of motivation. Motivation is all about getting started and consistently taking action, making sure we get back on track when we fall off the bandwagon. There are two types of motivation: intrinsic and extrinsic. When interest or enjoyment in an activity comes from within us, we experience intrinsic motivation. Intrinsic motivation associated with doing what you love, therefore, strongly correlates with sustained behavioral change and improved well-being because the activity itself brings pleasure. If you enjoy what you do, the activity and the goal will collide so that both your interest and experience of work are enhanced.

On the other hand, with extrinsic motivation, the outcome you desire is separate from the activity you engage in to achieve it, which will make dips in motivation more likely. Richard Ryan and Edward Deci highlighted the three innate psychological needs which must be satisfied to enhance self-motivation and mental health: competence, autonomy, and relatedness. If we feel competent in a behavior, either as a result of feedback, communication, or rewards, our intrinsic motivation will be greater. Relatedness is often more relevant to extrinsic motivation. If a behavior is valued by a manager, client, or friend, we will feel a sense of connectedness with them, which will lead to internalization of an extrinsic motivation.

It’s not just goals that are too hard that can cause demotivation. If a goal is too easy or will not lead to a suitable reward, you may lack the drive to pursue it despite it being achievable. Whether your motivation for change is internally or externally motivated, you cannot sustain motivation without having an aim clearly in focus. Motivation will only come once we have started a task or behavior, not before we get going. Rather than letting tasks accumulate until they feel insurmountable, the best strategy is to generate the momentum required to conquer a long-term goal by consistently showing up every day. As Confucius put it: “The man who moves mountains begins by carrying away small stones.”

To apply these concepts practically, here are three actionable pieces of advice:

1. Focus on the right goals. Intrinsic motivation will occur naturally if you choose a goal you care about. If you don’t feel committed or connected to the goal, you’ll need to rely mostly on willpower, which isn’t sustainable in the long-term. Take the time to reflect on what truly matters to you and align your goals accordingly.

2. Practice self-reflection. To promote motivation, you need to take care of yourself. Allow time for self-care, such as reading and exercise, and commit to self-reflection. Understand your values, strengths, and areas of improvement. By knowing yourself better, you can align your goals and actions with your true desires and motivations.

3. Embrace the power of consistency. Motivation is not a one-time event but an ongoing process. Consistency is key. Instead of relying on bursts of motivation, aim to show up every day and take small steps towards your goals. By consistently taking action, you build momentum and make progress, which fuels further motivation.

In conclusion, Google's PaLM sets a new benchmark for AI language models, demonstrating impressive performance and pushing the boundaries of what is possible. Meanwhile, understanding the science of motivation is crucial for achieving sustained behavioral change and personal growth. By harnessing intrinsic motivation, fulfilling psychological needs, and setting meaningful goals, we can cultivate motivation and work towards our aspirations. Remember to focus on the right goals, practice self-reflection, and embrace consistency in your journey towards success.