dw_arthur

Paradoxically, inflation has contributed to me taking a sabbatical. While I live in a LCOL area and made ~140k/year it just no longer felt worth it to work as I saw my retirement accounts start to match and exceed my salary in yearly gains. I do plan on going back to work in a part time manner, but inflation has killed any reason for me to work hard at a job for that level of salary. Furthermore, the feeling of "what's the point" around white collar work has never been more intense.

I assume they are working on low stakes software. Does it really matter if you use an LLM to code a scheduling app for a hair salon or veterinary clinic?

Sounds like internet addiction with new programs/features being the high you chase instead of new tweets, articles, or videos. It's beyond clear that we all need to be cutting back on our computer/screen usage.

I don't understand his audience still being in denial about what is coming. There may not be a job apocalypse but there is a decent chance tough times are ahead. Eventually, his audience may turn into outright Luddites. If that happens I sort of hope they don't stop at data centers/AI and go for the whole project of the Internet.

Good news, I was wrong. The discount only applies to Deepseek v4 PRO. Deepseek v4 Flash is not currently on discount which means the dirt cheap price will stay the same.

For my hobbyist purposes Deepseek v4 Flash has replaced Claude Code because I was also sick of hitting 5 hour limits with Claude. Right now, the only thing I miss from Claude is multi-modal image support. I can work around no image support since I can use v4 Flash all day and spend around $1. I am aware Deepseek is currently discounting their API at 75% off so I may try out another provider once the discount is gone at the end of the month.

At this point if feels like if you properly scope your work open weight LLMs are adequate.

A surveillance state was always inevitable once wireless networking, GPS, and cameras were ubiquitous. If you say this isn't true, show me anywhere in the world with these technologies that is not headed down this path.

Injecting unregulated chemicals into your body is something you may regret in the future. If you inject a Chinese peptide to improve your sleep quality score and three years later you're diagnosed with cancer the inscrutable nature of causality in the body may drive you mad.

We have been since we bound consumption to the internet. All of this was inevitable after that.

LLMs are powerful systems that eventually may be a requirement for being able to economically participate in a large portion of the economy. For this and other reasons it's important that people are able to control their own LLM.

Look at how much Google has changed over the years in the pursuit of profit. What will ChatGPT and Claude look like when they are pushed further down the profit maximization path?

Even two or three years ago I had ideas for projects but I could see the models were not ergonomic for my uses. I decided to wait for better models and sure enough the agentic models showed up which are much easier to use.

Next thing I'm waiting on is building a new server for a powerful locally hosted LLM in 5 years. No need to go through the headaches and cost of doing it now with models that may not be powerful enough.

Throwing an aside here that anyone interested in 1940s war technology must check out the old BBC documentary The Secret War (1977) which goes into depth on solving the engineering challenges of the war.

As someone who has done QA on white collar work it's tiring looking for little errors in work reports. Most people are not cut out for it.

It was okay writing in the context of marketing. A normal person never wrote like that.

Low information density is a big AI tell. Also sheer length is another AI tell. The LLM pukes out a bunch of text and the "author" either doesn't have the skills or energy to edit it.

It should show in decreased revenue for the company you didn't buy the product from. It also should show up at your company either as increased profit margin, increased investment, increase in total employee wages, or increased dividend payout.

If this is happening on a widespread basis in the economy we should see evidence of it sometime this year and that's what investors are anticipating with SaaS stocks.

It's funny that so many people are using AI and still hasn't really shown up in productivity numbers or product quality yet. I'm going to be really confused if this is still the case at the end of the year. A whole year of access to these latest agentic models has to produce visible economic changes or something is wrong.

Citrini charges like $1000/year for a subscription. It's ridiculous he is using AI generated charts.

As a long time computer hobbyist who grew up in MSDOS and now resides in Linux I'm starting to wonder if I am not more connected to computing than a lot of people employed in the field.

I assume if someone used an LLM to write for them that they must not be comfortabley familiar with their subject. Writing about something you know well tends to come easy and usually is enjoyable. Why would you use an LLM for that and how could you be okay with its output?

The Jevon's paradox example of plummeting costs to create videos doesn't make sense to me. If people are already watching 7 hours of videos a day how much more time do they have to consume video? There are only 24 hours in a day. Earth is only so large and can only handle so much pollution. There are limits we need to talk about here. I guess Elon hand waves it away by saying we'll go into space but that remains to be seen.

Everyone should have their own private evals for models. If I ask a question and a model flat out gets it wrong sometimes I will put it in my test questions bank.

Is the progress of LLMs moving up abstraction layers inevitable as they gather more data from each layer? First, we fed LLMs raw text and code and now they are gathering our interactions with the LLM regarding generated code. It seems like you could then use the interactions to make a LLM that is good at prompting and fixing another LLMs generated code. Then its on to the next abstraction layer.

I think a social norm of disclosing AI use at all times would be great. People and companies should also be held 100% accountable for anything created using AI.

I think the GPT3 or 4 minute mile moment for robotics will be when we see a robotic hand with the dexterity of a 6 year old. Once that happens it will quickly be over.

We've been building a turn key police state for the past 25 years.

This has been my experience with every sports of weightlifting injury I've had and I'm in my 40s. The body wants to heal or at least compensate in some way. Light activity is often better than rest. I've got a knee that is acting up a little bit but I think I've figured out how to keep it healthy while running.

Why publish anything for free on the internet if it's going to be scanned into some corporation's machine for their free use? I know artists who have stopped putting anything online. I imagine some programmers are questioning whether or not to continue with open source work too.

Sounds like philosophers will be the new programmers if playing around with language and definitions is all that will be left.

Reading this I feel the same sense of dread I get watching those highly choreographed Chinese holiday drone shows.