wuj

A good analogy is lifting. We lift to build strength, not because we need that extra strength to lift things in real life. There are plenty machinery to do that for us. But we do so for the sense of accomplishment of hitting our goals when we least expect it, seeing physical changes, and the feeling that we are getting healthier rather than chasing the utility benefits. If we perceive lifting as an utility, we realize its futile and meaningless. Instead, if we see it as a routine with positive externalities sprinkled on top, we feel a lot less pressured to do so.

As kelseyfrog commented already, the key is to focus on the action, not the target. Lifting is not just about hitting a number or getting bigger muscles (though they are great extrinsic motivators), its more of an action that we derive growth from. I have internalized the act of working out that those targets are baked into the unconscious. I don't overthink when I'm lifting. My unconscious take the lead, and I just follow. I enjoy seeing the results show up unexpectedly. It lets me grow without feeling the constant pressure of my conscious mind.

The lifting analogy can be applied to writing and other effortful pursuits. We write for the pleasure of reconciling internal conflicts and restoring order to our chaotic mind. Writing is the lifting of our mind. If we do it for comparison, then there's no point in lifting, or writing, or many other things we do after all our technological breakthroughs. Doing what we do is a means to an end, not the other way around.

The part on stories reminded me of Neil Postman's arguments in Amusing Ourselves to Death.

> Being told a story is to be infantilised, somewhat: to suspend one’s critical faculties.

One key argument in the book is that TV shows offer significantly less information density nor coherence as that of written mediums. They are optimized to reduce cognitive load. Thus our ability to think and process information diminishes greatly when there's nothing to think about. This is essentially what storytelling is - piecing together loosely related information to elicit an emotional response. The more harmful aspect is that they give us an illusion of learning, which the author also articulated with this quote:

> ‘The story wouldn’t be any good if you came back to your normal life completely unchanged, and having learned nothing, or having had no new observation,’ Vogler told me. ‘I think that we are always searching for upgrades, improvements in our behaviour, in our performance, in our relationships with other people.’ Films, he says, offer the opportunity for ‘slight improvement’.

TV shows wrap thin veils of lessons around stories. We feel like doing something fun and learning while learning something. Why not do more? So we consume more of it and spirals down into a self-reinforcing loop. But it is often not the case in real world. Learning is challenging. It's meant to confuse you and question your preexisting beliefs. We are numbing ourselves by associating what we watched in flashy media as concrete and substantial knowledge. The real takeaway is the experience and the emotional response from those dopamine-inducing flash cuts. When we associate learning and by extension, thinking, with emotions, that's when our critical thinking degrades and we become "infantalised."

Absolutely. However, as OpenAI forms partnerships and begins to offer ChatGPT as a plugin across various platforms, many of those niche applications would be replaced.

I think it will be less of a replacement and more like a partnership, per se. It will be hard for OpenAI to challenge services like Gmail due to the network effect. Same with Microsoft 365: People are used to that ecosystem. The success of partnerships hinges on whether Microsoft and Google can develop their in-house models and integrate them into their core products. OpenAI's partnership with Apple was a successful example of this strategy.

Part of Tumblr's downfall also comes from their change of stance on NSFW contents.

It starts to feel like a trend where OpenAI is integrating features that were previously implemented by GPT-wrapper startups into ChatGPT. While these startups have added value by enhancing user experience, the trajectory is leading towards an ecosystem where these functionalities seamlessly integrate. The future will be challenging for those startups.

My experience using GPT4-Turbo on math problems can be divided into three cases in terms of the prompt I use:

1. Text only prompt

2. Text + Image with supplemental data

3. Text + Image with redundant data

Case 1 generally performs the best. I also found that reasoning improves if I convert the equations into Latex form. The model is less prone to hallucinate when input data are formulaic and standardized.

Case 2 and 3 are more unpredictable. With a bit of prompt engineering, they may give out the right answer after a few attempts, but most of the time they make simple logical error that can be avoided easily. I also found that multimodal models tend to misinterpret the problem premise, even when all information are provided in the text prompt.

Time series data are inherently context sensitive, unlike natural languages which follow predictable grammar patterns. The patterns in time series data vary based on context. For example, flight data often show seasonal trends, while electric signals depend on the type of sensor used. There's also data that appear random, like stock data, though firms like Rentech manage to consistently find unlerlying alphas. Training a multivariate time series data would be challenging, but I don't see why not for specific applications.

This tool reminds me that the human body functions much like a black box. While physics can be modeled with equations and constraints, biology is inherently probabilistic and unpredictable. We verify the efficacy of a medicine by observing its outcomes: the medicine is the input, and the changes in symptoms are the output. However, we cannot model what happens in between, as we cannot definitively prove that the medicine affects only its intended targets. In many ways, much of what we understand about medicine is based on observing these black-box processes, and this tool helps to model that complexity.

All contact forms should have a feature similar to Google Form's "Send me a copy of my response" for recordkeeping.

That moment when you want to read a tweet just to be prompted to login.

Greenhouse and Lever have the most convenient job application interface IMO. The application area is one page, which means you can navigate using tab. There's also no need to create an account and verify email address (Though I understand why some portals do that to prevent spams).