crnkofe

I noticed its been trying to drag out the conversation on every topic. Maxxing token consumption for the dollars instead of staying on point. Makes sense for them but its frustrating for me as a user. I suspect this is going to be the AI version of enshittification. Now its only missing some badly needed in-stream ads.

I wonder if this will allow us to find backdoors in the models. Given the obscurity of the weights it shouldn't be too hard to hide spyware and backdoors inside your average LLM model.

Just learned about his position on security today. Been a Protonmail user for a while now. Initially I switched to detach myself from the US corp monopolies. Its just a matter of time now before I shutdown my Protonmail account. CEOs should understand that company ethics and security are not just black pixels on a white background.

I also started to just abandon writeups obviously written by LLMs. They feel (and usually are) just cheap marketing/selling posts. Not to mention why would I read stuff nobody wrote and likely never read themselves? There's little value in such content. I don't even point LLMs to such websites to get a TLDR. Part of why reading is interesting is the author, their history, beliefs and style. In the world of LLMs content, the author is some undefined diffuse alien with a made up personality.

AI has become such a hollow term. A literal tautology if you will. Everything and anything follows from AI. From devaluing education to human extinction. As time goes on the AI bros are even turning it into a dystopian fanstasy where finally there will be no need to interact with humans. Imo best way to deal with all the AI marketing BS and spam is to take a few deep breaths and head outside. Living in a state of fear an panic isn't going to result in any self-beneficial rational decision making.

I think there's a balance to be had. Leaving everything to agents when it comes to throwaway prototypes, one-off test env. scripts, extra tests I'm fine with. Doing the same with features - absolutely not - especially when it comes to financials. Seems like Meta is pressuring engineers into becoming expensive spec generators because they want to be in the cool kids club. Time will tell if that's a good play but to me it just sounds like a quick and dirty way to obliterate their engineering culture.

Every senior I've ever worked with at some point embraced the idea of minimalism. Produce minimal changes with a minimal surface. Worked well at startups and especially well on large spaghetti legacy projects. Seeing all of a sudden now how suddenly number of kLocs, PRs, token maxxing, number of tasks in progress is again a metric with which non-engineers brag about on engineering conferences feels like the industry has reverted to some braindead pre-infant state. I guess all mistakes need to be relearned at every hype cycle.

Having been working on a lot of low complexity code over the last few years I don't really sympathize with the author. I mainly use LLMs for the chores. And there are many many chores in software development. CRUD APIs, validation, yet another DB service, yet another script, minor version upgrades etc. Chores are also very straightforward to review. I do reserve the time to do normal engineering especially for the rare interesting algo and don't get this obsession with how everything must be agentic at all costs. Honestly seems like crypto propaganda all over again. If productivity "really" has gotten 10x doing some tasks without LLMs won't even put a dent in your average dev. output. What I do grieve is a lack of non-AI news. Might just stop following HN for a year or two and wait for the LLM spam to die off.

I tried some of these note taking apps but eventually just ended up with a self-hosted wiki + Git/Forgejo/markdown. Once I had enough material, diy projects, scripts, random media just thinking about moving them around, backing them up, sorting them resulted in a heavy mental strain. Having a simple, familiar system to work with just brings a lot of value.

I find LLMs are a hit&miss. Can be great at rewriting a function from lang A to lang B. Fails completely at a refactor. Great at coming up with a bunch of networking hosts/IPs to test a particular func. Fails when doing simple validation. Great for prototyping an alternative UI but not even remotely production ready. Always confident regardless of whether the result is correct or false. Run it three times with same query get 3 different results etc. Plenty of claims online how people have become a 1000x developer but 0 actual examples of working code in production. Given they're stochastic by design all of this makes sense in a way.

There's no clear path moving forward. Overreliance on LLMs means your knowledge will exponentially decay and you will absolutely crash any future tech interviews becoming unemployable. Not using it for some quick wins feels wasteful. Finding balance between the two extremes in addition to all existing software development woes is really hard.

I used to be happy to have one conclusion derived from a meeting. Now with every word catalogued via AI summaries and pages of slop you'd think we have gotten somehow more productive. Alas the amount of conclusions didn't change. It does look like a lot of work was produced though so maybe on the plus side I can "prove" we did talk about stuff when eventually management comes looking for answers.

AI-generated style right now has a what I'd call a grotesque aesthetic. Its offputting and immediately recognizable. The biggest offender being the rounded fonts and "wrong" smiles. Its probably going to get better but until it does I will feel free to ignore it same way as highway commercials.

Software engineers are splitting into two camps either by choice or by being forced to by management. Those who have the self-discipline to use LLMs responsibly and those that are going all-in on AI. Imo there's a place for both camps but I don't think they can coexist in the same team.

The language of these LLM posts makes me think they're considering the option to be a defense subsidiary. It doesn't seem like they're actually constraining or limiting the models in any way. They don't seem to understand how the model actually works. Poke the beast and see what happens. Also the use of passive language as-in AI is becoming more and more of a threat as opposed to the reality where they're making the model more and more aggressive and useful for military is very hypocritical.

Lots of people mention this higher abstraction level but its not really the case. C has a grammar. You write C and get C as output. Assembly is hidden from the user. Natural language also has a grammar ("sort of"). You write prose into your preferred cli/app of choice BUT it outputs code in C or other language. So its basically a code generator. A code generator is not even remotely a form of abstraction. Its just a facade that breaks down as soon as the generator fails to produce a meaningful result.