bazera

What's your day-to-day diet like? Like for 80-90% of the times, what do you eat? I try to lose weight while building muscle and here's what I eat: protein pancakes (protein powder, banana, egg) chicken breast + rice / buckwheat eggs cottage cheese tuna salad some fruit That's it basically. 1900-2000 calories and 160g protein. And on the weekends I might switch and bring in some variety, but still stay in these caloric range. What about you?

@Natasha Tori Maru Yeah, all that makes sense in the construction industry. I think it will stay as LLM for a long time to come, so you should be safe. I think there aren't many industries that LLMs affect that much, except for maybe software and couple others. And even there, it affects mostly positively for experienced people. Aside to your work, do you use it for your personal life in some way? It's pretty useful when you are learning new things for example. One interesting usage that I've found was the following kind of questions: And since it has memory form all your past conversations, it might give you some answer that changes your perspective or gives you new insights.

@Stick Yes but only if you steam the vegetables.

I fear this will happen not only to soft. devs but most people across the board. I hope I'm wrong

@Lyubov Yes, same. Again, same for me. If I didn't know what I'm doing, it would all turn into a mess quickly. Opus 4.6 model got refined and does much much better job at coding for example. Recently it saved me a huge a amount of time, probably weeks, in one of the tasks I had which was about migrating an old design system to a newest version where they had breaking changes. Without Claude, I'd go through hell probably. All that being said, the only reason I was able to use it effectively was that I had years of manual experience. I think people who say that LLMs will replace employees, haven't actually done much work with the assistance of LLMs.

It will make them even lazier and depended on some chatbot.

https://www.instagram.com/reel/DVgNBb9kymA/?igsh=cjdsYWlreXc2anZi What could go wrong

Yesterday was a shitty day. The junk food that I ate the other day caused stomach issues and I didn't do much except for laying on my bed in struggle. Today I feel better.

@VioletFlame If that happens, not having ChatGPT is the last thing we'd worry about But yeah, humanity is getting very depended on thinking with LLMs. Me included. I'm in the process of learning how to balance independent vs LLM assisted thinking. Because it's very tempting when you've got some question to just ask Claude. But it's not responsible way of using mind for sure.

My workflow depends heaviliy on AI (Claude Opus 4.6), but it always needs proper guidance, reviewing each line it produces, often makes something I don't want and I'm guiding it to course-correct. Sometimes it just can't fix the issue, even when it sees the full context. That's when I step in manually and help it myself. It's a whole process where me and opus are both heavily invested. I'm not sure why are we discussing human replacement when it's so apparent that even best LLM models need human involvement and orchestration. Can LLMs literally replace you and do your job? No (depends on what you do) But it changes the workflow for sure. With the obvious downsides like the possibility of skill degradation, mind rotting, no more incentive for technical debth, etc. And it's not fun anymore. When I was coding in the past it was more adventurous, I felt I was advancing my skills daily and weekly. Now I just talk with a chatbot and review the code, and occasionally do debugging manually. Sometimes I wish we could go back. But on the other hand, it can super charge the learning process. LLMs can be used very effectively to assist your education. It doesn't replace books but supercharges them. When reading a book simultaniously I can chat with LLM about a topic I didnt fully understood in a book and get it cleared.

The trick that I use sometimes is that if you set phone orientation on landscape, you'll get desktop version formatting options. I use what I want and then switch back to portrait mode. Takes extra time but better then not having options at all.

@integral While all that being true in my experience as well, Antropic CEO's claims are still delusional. AI won't iteratively improve itself in 6-12 month time to the point of making developers obsolete. He is not going to replace his developers in 6-12 month times with opus. It can't just code C compiler and fully-fledged Figma clone why we sit and sip coffee. Still requires very active developer involvement. It's just that we don't manually type most of the code anymore.

If somebody told you yes, would you believe it? Some people who astral project claim to meet their deceased loved ones. I haven't done it so I can't verify.

You guys with your kangaroos and wild pigs, and I'm here eating chicken breast weeks in a row. I need to step up my meat game

Another good one, just dropped. Cal created a segment of his show called "AI Reality Check". He'll be doing it each week. We really want this kind of information now.

This is really far fetched man. So you are saying that some combination of opus agents can create a better version of Invision board? No, I don't think so.

Do you mean something akin to what Ken Wilber would call structures (egocentric, ethnocentric, worldcentric, integral)? You do not perceive structures as objects. They appear as patterns in how experience is organized: what you notice, what you ignore, what feels meaningful, what explanations your mind generates. But it isn't found in senses. They show up as automatic interpretations, conceptual frameworks, sense of identity, how complex your understanding of situations is. Do you percieve morality the same way? They have certain reality but are those ultimately true? Like, in all circumstances?

It's especially problematic when it comes to new hyped up techonologies that we don't really know how it all works under the hood (LLMs for example), and that makes us more gullible to their confidence. And giving authority away to confidence really re-shapes the worldview. For example, if you believed what he says that in 6-12 months software will write itself, it would affect your life decisions in drastic ways if you are related to software industry in some way. We should be careful in what we believe in. But as you said, amidst the flood of notifications and decesion fatigue people aren't used to doing that.

I really want to meet him in person and study at least couple weeks under his guidence while he's still around, I think it will be a very insightful experience for me. I really envy Americans who live there and can attend his workshops, like, one of the top reasons that I wanna get US Visa (not sure if that works out) is that I get to interact with Peter, not putting him on pedestal or anything, it's just so interesting to interact with a person who takes truth that seriously. The new book is very intriguing.

Some thoughts on this video: https://www.actualized.org/insights/more-ai-delusions My god, they bullshit us with such confidence and straight face. When what that Anthropic CEO said will definitely not happen, that in 6-12 months Claude will do perferct job at software development, will there be anyone that makes him accountable to that words? Or everyone just ignores it and allow him to bullshit us again with yet another 6-12 month fearmongering claims. And the fact that they straight up lie in the C compiler ad, is just crazy. I was more sympathetic towards that company, because I really like their tools, but this just makes me certain that I shouldn't believe anything that they're marketing so deceptively. It's a good example of how important it is to do independent research into a thing instead of just believing a marketing ad.

It gets less annoying after a while Once you get enough experience with writing pure css, then it actually becomes fun and challenging to work on difficult layouts. It's always useful to know the fundamentals when you use some abstraction like Materialize, Tailwind or Boostrap. It will actually make you more effective.

Some insights: I have to plan ahead, what comes after another, and limit time in-between, especially till 8pm. After that I can have more leeway. I don't really like structured day, but this is what I need to do at this stage of my life to succeed in what I'm trying to do. Most important thing for me to do is removing cheap dopamine habits (PMO, junk food, doom-scrolling, etc). If I don't do this, I'm not able to build other healthy habits (spiritual practices, exercise, weigh-loss, consistent study and read routines, etc). If I actually do it, all these things become much easier to maintain. Overview of what I'm trying to do at the moment (all is important, but sorted by most important to least important) [HARDEST & MOST IMPORTANT] Removing / managing in a healthy way cheap dopamine habits Losing excess 30kg (66 pounds) Building muscle and consistent exercise regime (weight, running) Gaining momentum in foundational spiritual practices (Yoga, Meditation) Building daily reading and study regime (1.5h daily) That's it basically. How I want to structure my day: Wake up around 7am Meditation practices first thing in the morning Working, if there are some leftover tasks from yesterday, or else exercise Cooking Study 1.5h Work and reading 1.5h in between or in the evening That's it.

Okay, I relapsed with PMO, and all my other structure crumbled because of that. This is basically the reason why I'm stuck at the current situation for the last 10 years. I knew all this shit 10 years ago, I knew the benefits of meditative practices, excercising, healthy-eating, consistent deliberate practice in some craft over time, etc, but I kept myself into the same pattern of relapse => blame, guilt => crashing down all the structures => next trial => relapse, etc. This won't stop happening until I don't manage this habit. I'm not sure if anybody actually read this, but this will be my public commitement for this. Because if I don't handle it, nothing else will work. Everything that I wrote here as plans, none of it. It's that serious. If I take self-responsibility seriously, and not just in theory, I have to manage this part of my life because it's destructing for all the other parts and keeps me stuck for years. I have to be really responsible here. Not in theory, ACTUALLY.

Me too. AI could really boost that process, it could assist you in improving your solutions, give you new perspectives, help you when you're really stuck and just wasting time with no hope. The key word in "AI assistant" is assistant. I'm trying to use it as assistant, not just for solutions generator. So it's not really an AI issue, it's more like a people issue. People are irresponsible, and giving irresponsible people these tools without proper guidance might go sideways.

@Joshe Sorry if I misinterpreted. I'm subscribed to his channel and watched that one when he posted it, he makes good points. This is exactly my experience as well. That might be the case, but that doesn't mean it will be sustainable long term. Also, "making money" isn't a good indicator to me for quality software. Slop can make money as well. The project that I'm working on has been in production for 8 years as of now. It takes a really careful usage of AI to not mess it up over time. It won't be apparent that you're messing up the project with 1-2 promtps. It will be a slow process. If you let AI coding assistance run amok only with people with little technical knowledge on project like that, it will become disaster in couple months. Maybe it can work for smaller scale projects.