Hacker Newsnew | past | comments | ask | show | jobs | submit | __0x01's commentslogin

Often when I look closely at the output of LLM generated code, I see repetition, redundant logic and deeply hidden bugs.

Notwithstanding the above, to my understanding LLM services are currently being sold below cost.

If all of the above is true, at some point the degredation of quality in codebases that use these tools will be too expensive to ignore.


These LLM tools appear to have an unprecedented amount of access to the file systems of their users.

Is this correct, and if so do we need to be concerned about user privacy and security?


We should be absolutely terrified about the amount of access these things have to users systems. Of course there is advice to use a sandbox but there are stupid people out there (I'm one of them) who disregard this advice because it's too cumbersome, so Claude is being run in yolo mode, on the same machine that has access access to bank accounts, insurance, password manager and crypto private keys.

My understanding was that atherosclerotic plaques are comprised of cholesterol or fatty deposits [1] and that these can lead to CVD.

The fat mechanism I understand, but what is the mechanism for sugar in CVD?

[1] https://www.heart.org/en/health-topics/cholesterol/about-cho...


CVD requires a bunch of events to happen in sequence, I always felt like it was a combination of risk factors + luck that make a heart attack or aneurysm happen.

1. High blood pressure damages walls of arteries and veins

2. LDL Cholesterol gets into the damaged walls

3. LDL gets oxidized

4. White blood cells engulf oxidized LDL and form plaques

5. Hardened plaques chill, they are bad but not deadly, if a plaque breaks off you are probably dead.

Sugar is gonna contributes to 1 - 3, especially 3 it seems way more guilty of than fat. The one big thing that opened my eyes was that most of the LDL you get is going to be produced by your own liver. Regulating how the liver produces it is going to have a bigger impact than directly eating less/more of it.

It is kind of a luck thing though, you could eat like shit and never have all the events occur just due to dumb luck, or you could be a fit 45 year old and for whatever reason you get a plaque that breaks off and you aneurysm and die.


And the liver produces triglycerides from fructose which is half of sugar.

Consuming cholesterol doesn't normally change the level of cholesterol in your bloodstream - it simply leads to your body producing less cholesterol. Unless you're consuming gigantic amounts, or have some problems with your cholesterol regulation, dietary cholesterol is completely safe. It's only if your blood work shows elevated cholesterol levels that you need to start paying attention to cholesterol intake. This is in fact very similar to what happens to blood sugar levels, in fact.

Pretty much every health authority will tell you that high blood sugar damages blood vessels, thereby enabling the formation of said plagues.

Healthy adults consuming some dietary sugar doesn't cause persistent high blood sugar, though. That's diabetes.

it's not just sugar. https://en.wikipedia.org/wiki/Glycemic_index#Grouping

all simple carbs are the devil, but we can't possibly feed billions of people actually healthy food - organic vegetables, nuts, and animal products, so come drink your corn syrup.


The sugar industry (topic of this article) can only be blamed for sugar, though -- not all high-GI foods.

And you can replace "sugar" in what I said earlier with "high-GI foods" and it doesn't change a thing. Persistent high blood sugar is diabetes; it isn't dietary.


>Persistent high blood sugar is diabetes; it isn't dietary.

how is it not dietary if consuming most carbs spikes your blood sugar for hours, which, with three meals + snacks + starbucks slurry, means elevated blood sugar 20+ hours a day?


It doesn't happen in non-diabetic people. It's different in type 2 diabetics who will see large swings in blood fat and glucose after meals.

Please can you provide a source for the above?

Sugar causes inflammation, and inflammation damages arteries. It is this damage that then leads to accumulation of fatty deposits, as damaged arteries basically lose the protective layer (think of equivalent to a non-stick coating). But that doesn't mean dietary fat is what actually caused the plaque.

Poor dental health also contributes and nothing pushes poor dental healthy like a high sugar diet.

Engineering Room, panning over a bunch of hot Blackwells

"I can't change the laws of physics!"


The monster babbleth no more, sire.


Is cost a risk here? I'm assuming that sometime in the future the price for vibe coding/engineering will go up significantly.


It already is. I saw a demo of someone spending $30 an hour on Cline to do something that would be achieved better, for free, and faster, with a framework.

That's more than $5,000 a month assuming 40 hours a week.

The end result didn't compile by the way.


I have noticed that many of the most used and loved projects use C/C++.

With all of the unpopular press that they get, why has history often proven C/C++ to be the right choice time and time again?


The one thing I really regret is listening to the kind of opinionated people that will tell how a certain popular language is bad and you shouldn't use it.

Twenty years ago when I learned Python, people told me it was a waste of time. Sure it is a nice language but you will never get a job doing it.

When I got into PHP ten years ago people told me it is a horrible language and basically dead already. Still, PHP was key to my career and I am making good money programming with in in 2025.

I used to be a C++ hater because that was the cool thing to do. Then I tried Unreal Engine and had to use C++ and discovered it is... fine. Really. There is a good reason it is heavily used in the gaming industry. I don't totally love it. The compile times are kind of annoying but C++ is not the only language suffering from that. If you need the performance and the ecosystem, C++ still isn't the worst choice.


There are only two kinds of languages: the ones people complain about and the ones nobody uses.

-- Bjarne Stroustrup, The C++ Programming Language


The most used Desktop projects were started before 2015, around which new languages were stable enough to consider app building. Those apps fill their niche well and people don't need to start new projects for these niches, and new developers add to those projects instead of replacing them, which is a good thing. The cloud, AI, mobile (or crypto ..) by example are another story.


Part of the answer is in term C/C++ itself. C is a very popular language because the history around Unix and FFI. C++ builds on that popularity by being sort of backwards compatible with it and being what Microsoft Visual C++ supports if you want more languages features beyond ISO C90. Which brings to mind another factor, that people used to pay for compilers and IDEs. The hobbyist had to consider price as well as functionality, portability, performance and popularity in what programming languages they choose. Given there were several C++ compilers to choose from and the popularity of Windows it makes sense that C++ won.

Why has nothing replaced it? That depends on what you mean by replaced. Python, and Typescript/JavaScript have replaced it in many places. It's just low level programming where C++ has yet to be superceded. For that kind of programming there had not been many that could even approach the space until LLVM-based languages started coming out recently.

Some things come down to the OS system. We're still using the C FFI and other system elements designed around C. So until something better replaces those we're still using C on some level.


Long term stability is pretty important and C/C++ has a pretty good record of that. A lot of the bad press is also for ideological reasons, not entirely practical ones.

Finally, I think it's because the C/C++ manages to stay hidden. You don't need to know that it's written in C/C++ to use it.


Survivorship bias


Don’t trust hype.

”Unpopular press” generally means ”non-expert clickbait opinions”.

C++ as a language is abominable but the ecosystem (compiler support, tooling, libraries, established knowledge) is very hard to beat.


Is this cheaper to run than the gas equivalent?


What LLM does Phoenix.new use?


claude 4 sonnet as the main driver atm, and a mix of smaller models depending on the scenario


I'm building an agent right now (or rather, extending an agent I build a few weeks ago in a couple hours) and I would love to hear more about which scenarios get assigned to which models.

(I almost just asked you on company Slack but figured the answer would be more broadly interesting.)


Based on the blog post, Phoenix.new uses Claude 3 Opus as its underlying model, which explains its strong performance with Elixir/Phoenix codebases.


I thought this was going to be an essay on the impact of English, Math and Programming on humanity.

I would place English (or all spoken/written languages in general) first, Math (as discovered) second, and programming languages last.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: