HNHacker News
TopNewBestAskShowJobs

lnenad

981 karma · joined November 2, 2018

Eng Mng with 10+ years of exp; currently working with AI/LLMs

nenadspp gmail.com

grafly.io

logdot.io

getlipwise.com

submissionscomments
lnenad··on The AirPods Effect
On a gut level I agree 100% as I'm feeling the same feelings and prefer no contact. But on a biological level it's definitely better to have more social interactions and talking to strangers specifically has many benefits to both you as an individual and for creating a better society.
lnenad··on Ask HN: What are you working on? (June 2026)
Of course. Primarily the styling and a lack of symbols tldraw/excalidraw, and for draw.io I dislike the UI/look and feel.
lnenad··on Ask HN: What are you working on? (June 2026)
I didn't like how diagram making apps were built so I developed https://grafly.io

I also vibecoded/built https://github.com/lnenad/difiko as AI generates a lot of code that needs a nice way to review it.

I'm also in the process of preparing a soft release for http://logdot.io as the market has no access to low friction logging/observability platforms.

lnenad··on Why AI hasn't replaced software engineers, and won't
The world isn't US based rates. There are plenty of people that would be taking these "non-economical" projects and being extremely happy with the money they can make from them.
lnenad··on Why AI hasn't replaced software engineers, and won't
I'm talking about actual live services that deliver money to someone, not toy projects. Do you think there's even a slight percentage of those or are 100% of the projects hosted by these services throwaway shit?
lnenad··on Why AI hasn't replaced software engineers, and won't
Please try to read my message again. I never said the things you're implying I said. I literally said not gonna replace all, and low effort low risk stuff.

Do you think that not even 0.00001 projects on those websites could have been a good payday for a software engineer/team? Do you think what took 3 people before for a low effort saas is not going to be done by 1 person now?

lnenad··on Why AI hasn't replaced software engineers, and won't
It literally has and will even more in the future. It won't replace *all* software engineers but once the genie is out low effort low risk stuff will be done by an AI. Loveable and such have so many live projects, the alternative was a human building those.
lnenad··on Ask HN: Did Anthropic Just Win?
It's not a valid point imho as your paying for tokens, not hours, per token value is much more important and the purpose of getting more lower value tokens does not exist (unless for a fraction of the price to get to the same result). But it's fine if you feel differently.

Yeah, I'm rooting for this otherwise we're in for a rough ride in the next couple of months.

lnenad··on Ask HN: Did Anthropic Just Win?
I'm saying it's not expensive tea if you have to drink more of the cheap tea to get the ~same effect.

That's why I'm posting this, my expectation (based on previous behavior) is that there should be a response from OpenAI. Considering it's missing I'm worried there's some secret sauce that others are missing.

lnenad··on Ask HN: Did Anthropic Just Win?
But even if 10x the price the value is there. Fable needs to loop less, spending less tokens, and is going to be much more correct so the price increase isn't a straightforward evaluation.
lnenad··on Apache Burr: Build reliable AI agents and applications
Claude Opus really loves this template when building websites. It's very funny how many times I've seen it for recent launches.
lnenad··on GPT-2: Too Dangerous To Release (2019)
You'll never make me wear a mask!!11
lnenad··on GPT-2: Too Dangerous To Release (2019)
Feels like a hundred years ago.
lnenad··on FrontierCode
> Total token consumption would also be another thing to consider as well, to rule out TPS.

There is a chart that compares this in the article.

> You should compare apples to apples. Weight them in a way that factors in total task completion time as the measure of "effort", not the arbitrary effort settings provided by the AI company. I don't care what the underlying effort level is, I care which model out of multiple, if running for the same amount of time, completes my task to a more accurate degree.

That's your opinion, my goal is exactly what this benchmark measures, the end result being something I can merge into the codebase based on some configuration setup provided by the lab. I don't run 50 agents in parallel and I am able to use the $100 Anthropic plan just enough that I don't go over the limits.

Also what is your specific argument into the benchmark findings considering that some problems are solved by Opus and not solved by Codex? Whether one uses more tokens or not is a completely different metric.

lnenad··on Ask HN: What are tools you have made for yourself since the advent of AI?
I didn't like how diagram making apps were built so I developed https://grafly.io

I also built https://github.com/lnenad/difiko as AI generates a lot of code that needs a nice way to review it.

lnenad··on I'm skeptical about efforts to revolutionize schooling
The problem with any such statement is framing and agenda. I agree with the premise that education has a long way to go to be universally good. But compared to undeveloped rural areas where people don't have access to education? Is it better? Everything points to an unequivocal yes.

I personally avoid giving money to people who are easily obviously having a net negative effect on society.

lnenad··on I'm skeptical about efforts to revolutionize schooling
I strongly disagree (having not read the book and having my own biases). History teaches us that any education is better than no education. Formal education is very far from just booklernin', you also learn *how to learn*, you develop social skills, you network, you start understanding how to deal with administration stuff, how to deal with failures and many other skills...

No country has perfect education, but what we call education is very useful for becoming a society participating human.

lnenad··on I'm skeptical about efforts to revolutionize schooling
I cannot imagine coming to a place that has a reputation of having higher educated people, and presenting that education is a waste of time. All research has some levels of agenda behind it, and it would be very easy to steer anything the author presents into a direction he wishes. On the contrary, literally every country on earth benefited greatly from education and the results are very visible in all aspects of human development (HDI, GDP, pollution, literacy, etc...), most recently China, and even more recently India. Both had huge numbers of illiterate people, compared to both advancing at an incredible pace when a large emphasis was put on education.
lnenad··on I'm skeptical about efforts to revolutionize schooling
> "Kids will learn anything that gives them social standing or self-worth in another way, whatever it takes to be a cool kid."

"nerds" would disagree with you. I think the point that OP was trying to make is the three groups are dynamic based on the topic. So the groups are not the same for biology, maths and for literature.

lnenad··on Gemma 4 12B: A unified, encoder-free multimodal model
Handy is open source and works flawlessly for me with Parakeet v3.
lnenad··on Alphabet announces $80B equity capital raise to expand AI infra and compute
I think the point OP was trying to make is when a person "buys low sells high" they pay taxes on the gains.
lnenad··on Microsoft builds MacBook Pro rival with NVIDIA-powered Surface Laptop Ultra
I mean, yeah ok? Not sure what your point is? This isn't directly related to this machine though. First Nvidia SoC and a much better vertical integration with now non trivial amount of experience in hardware stands to offer something that wasn't avaliable before.
lnenad··on Microsoft builds MacBook Pro rival with NVIDIA-powered Surface Laptop Ultra
It's not a valid comparison. Up until this (theoretical) machine, the playing field wasn't equal. Windows laptops in general couldn't really compare in many aspects with Macs since M1 (outside of gaming), batteries were horrible due to bad efficiency, performance wasn't amazing even with the best SKUs, and the distance between the vendor and Microsoft was always impacting different aspects of the finished product. Even with the ARM Surface, the ecosystem still wasn't ready for ARM, the performance was lacking. If this device doesn't cost an arm and a leg, it will offer something that is really the first instance of a Windows device that's a better choice than an M-Macbook in many ways (at least on paper).
lnenad··on Sony Launches Bravia 9 II and Bravia 7 II with 'True RGB'
I don't know what "early" represents for you so cannot comment on that but my close to 10 year old 55 inch tv still doesn't have it.

Either way there's something about OLEDs that my eyes prefer, I cannot verbalize it. Maybe it's placebo :)

lnenad··on Sony Launches Bravia 9 II and Bravia 7 II with 'True RGB'
Having had OLEDs for a while now and never experiencing burn in even in my monitor I believe this issue is either very overblown by the small amount of bad experiences and requires a large amount of misuse before it becomes a problem.
lnenad··on Ferrari Luce
I hate 20 inch, floating, glued to the dash tablets with such a passion. It cannot be such a huge monetary difference to have physical switches for the AC compared to this attention grabbing accident causing contraption that was never meant to be put in a human commandeered vehicle.
lnenad··on BambuStudio has been violating PrusaSlicer AGPL license since their fork
I mean, if you want to make your point, yes. But I think it would be logical to assume "US defence manufacturers" wouldn't be using Bambu from the start, regardless of what their track record is.
lnenad··on BambuStudio has been violating PrusaSlicer AGPL license since their fork
I think you're overestimate the value of these prototypes. The print itself is either a plastic render of the final product without any value, or it's a shell without any actual useful parts/machinery. If we imagine we're talking about the 1% of 1% of 1% which could end up as useful IP stuff, but which might be very hard/impossible to find/understand/do anything with it, for which cases don't use bambu.
lnenad··on Blog ran on Ubuntu 16.04 for 10 years. I migrated it to FreeBSD
I love people that aren't afraid to experiment and learn. As someone that hasn't had a formal education in software engineering (just in other kind of engineering) I learned the most by doing and failing.
lnenad··on Why senior developers fail to communicate their expertise
What are we talking about? Philosophically yes. Factually, no. In the context of a system innovation could be switching from one form that renders in 1 second to another that renders in 50ms. Stability isn't part of that equation.
← PreviousPage 5 of 15Next →