Comment: Evening rambles through the day and work
Transcript: Another thing for today was work stuff and, well, started off being in llama.cpp land again, just kind of helping out with some of the GitHub issues and making sure the code is working as expected, and it is, as far as I can tell, which is cool. So that's what I did after I went climbing, I pretty much immediately took a look at that and started working on that. I did a fair bit of reading today, more of Michael Pollan's book on building his house. I probably will read a bit more of that before I go to sleep today. And also, in terms of getting things done for the project, I implemented kind of the saving kind of function, and I still want to think through this one more clearly, but more or less, there's like a solid calf for things that you've already transformed. So basically, once you've labeled something as like, this is related to food, you never have to go back and do that inference again. And I think this will be probably pretty helpful moving forward. And we'll kind of see how important this is. And my guess is, I mean, I want to run this on things going in. But yeah, it's just a curious thing, basically. So yeah, that was what got done today. And tomorrow, I think kind of the big thing is, instead of doing it for each, probably doing it for all. Probably just a basic caching mechanism here, and being able to set that on the front end or whoever's querying it, how long to cache that for. And it probably won't be perfect to start, but that's okay. I might also try to make the cache kind of like a generic adapter. So we make a folder of caches and kind of do that kind of thing. There are also some questions around the code generation. I think the code generation needs to be dramatically improved. I wonder if breaking it down into further steps may help. But that's something I want to kind of visit soon. I don't know if it's going to happen yet. But I just want to clean up the kind of code that is being generated and go from there. And ultimately, I mean, all of these generations and everything, I think I'd like to be saving them in some way, shape, or form. So I really need to think about some logging system in a very real sense of what transformations are happening. When are they being regenerated? These kind of things to get a sense of what is training data? What needs to be trained? What needs to be fixed? Counting a bunch of metrics. Anyway, all of this is for later. I think for now, getting these basic things done will dramatically improve just the experience of being able to build these little components. And I think that's the main thing I want to do is just trying to close that gap a bit and make sure that that development experience isn't very painful. There are some quirks to it. But on the whole, I think that it's pretty important to try to nail this down. And then there's some other things to think about in terms of like, well, how do we do like additive transformations or transformations through time? And just trying to make sense of it in any kind of way that will work both in computer land as well as, I guess, more from the perspective of human as well and trying to collapse that gap. But on the whole, I think I'm just happy to get the caching thing done. It was like kind of a challenge just to get it done. To reason about like where to start. in terms of do the simple thing or do the more complex thing and solve that. And I just had to ultimately decide on doing the simple thing. A few competing things in my head going on at the same time. So, mostly regarding money. So that was distracting me quite a bit today. And just like, do I wanna start making things for Vision Pro? And thinking about those things in AR and being very, very interested in like, legitimately the friend's photo frame. I'm so curious to have that.
89.21% similar
Today has been a challenging day, with the speaker feeling overwhelmed by logistical tasks like taxes and job inquiries. They had a productive conversation with Danny but were left feeling aware of the amount of work ahead. They are also pondering ways to make money and considering the potential of experimenting with data and language models. The speaker is interested in the concept of "brain twin" and is curious about using it in a group setting with others, possibly collaborating with someone named John.
88.30% similar
The speaker had an eventful day, pondering about the challenges of data input while considering tab as a better alternative. They also reflected on the importance of finding balance and engaging in outdoor activities. After accidentally spilling water on the couch, they spent the rest of the day sewing and improving their skills. They mentioned the completion of their mom's sweatshirt and the priority of finishing their dad's garment. The speaker also expressed anticipation for a call with Dave the following day. The speaker begins by expressing uncertainty about the need for assistance and the importance of finding a solution, and then transitions into a discussion about their activities, including sewing and debating whether to go climbing. They also mention interactions with others, including a visitor and a call with someone working on a home assistant project. They express curiosity about artificial intelligence and self-reflection about their choices and the validity of their concerns. The underlying theme revolves around the need for accessible and contextual data and the desire to visualize and understand their thoughts and emotions. The speaker is pondering the potential of representing human thought in a computer and visualizing data. They question the user experience of digging into menus to see past interactions, suggesting that simple reminders may be more effective. They also explore the visual representation of a person by AI, wondering about its accuracy. Additionally, they express a personal interest in statistical information, such as the number of days spent with a specific individual. Lastly, they mention their intention to read the Pure Aloha Oath, emphasizing the importance of living with pure Aloha in all thoughts and actions. The speaker commits to embodying love and compassion in all their interactions with others, seeing everyone as part of a connected global community. They aspire to achieve inner peace and happiness by living each moment with unconditional love and open-heartedness. While acknowledging the challenges of consistently living by this oath, they are determined to strive towards being a better, more thoughtful, caring, and kind person. The speaker reflects on the practical aspects of their life, such as cleaning up a mess and organizing their home, and expresses frustration about technological complexities and the desire to track and visualize their daily activities. The speaker discusses potential job opportunities and muses about conversing with an AI that replicates their own brain, similar to the movie "Her." They express a desire to be part of something worthwhile and mention needing to respond to someone named Chroma. They contemplate talking to themselves and sorting through their thoughts, feeling enthusiastic about the subjects they want to share with the world, such as photolithography. Additionally, they reference finding better platforms for sharing information and react to a message from someone they contacted about acquiring a slackline. The text seems to be a conversation or stream of consciousness, with repetition of names and phrases. The speaker expresses uncertainty about what to say in response to a message, indicating a lack of understanding. The speaker seems to contemplate keeping in touch and checking in on the new year. However, there is an overall sense of confusion and difficulty in articulating thoughts. Dr. Lingonberry, an individual's friend, has been absent and has developed progressive memory loss. The cause of this is due to a diagnosis of a brain tumor. This diagnosis has led to both consequences of memory impairment and the presence of a brain tumor. These combined factors have significantly impacted Dr. Lingonberry's health and well-being.
88.11% similar
I woke up at 6 and got straight to work on my Llama.cpp, made some changes, and submitted them for review, hoping for acceptance and a merge. After work, I'm heading to Tourmaline to surf despite the mild rainy weather. Upon returning, I plan to relax, possibly finish Kyle's sweatshirt, meditate, do some shopping, and clean up my code, which I currently consider subpar. I intend to tackle some easy tasks on my to-do list to smooth out any issues and improve my code's overall quality.
88.07% similar
I'm making good progress on the AI grant application, with both the longer description and one-sentence summary feeling satisfactorily crafted. Despite some reservations, the video I've made is likely sufficient, and with most steps completed, I'm now moving onto the demo, aiming to showcase everything in a concise three-minute presentation. This will highlight one personal frame through which to view data, particularly emphasizing social connections and convenience in planning. My ultimate goal is to demonstrate the simplicity of asking a question to retrieve information and to focus on the two key APIs, store and query, to power the application. Focus around these two queries, as with them, essentially any application can be constructed.
The realization of the value in this application lies in its ability to perform searches quickly, efficiently, and accurately. There are multiple approaches to enhance its functionality, with a focus on both data storage and the improvement of search capabilities, which is currently the most critical yet challenging aspect. Concerns exist about the app's method of aggregating all processed data, which feels inherently flawed, though it's being temporarily accepted for the valuable data it provides. This tension between a recognized need for development against the reluctance to proceed with an imperfect solution underscores the complexity of the problem at hand.