mrburton
mrburton on Hacker News
Likewise, I this direction gives engineers a way also to fight back with all the hackers/scammers. It'll create so many new companies as well.
Great clarifying question; sorry I should have been more clear.
I mean working with open weights and in the future, creating their own open source models from the ground up. At the moment, that's not reasonable, but I'm leaving the door open for when/if there's a break through in training that greatly reduces the costs of training new models.
Right now, LoRA can give amazing results with small models.
Posted by
mrburton on Tuesday October 27, 2026 @02:50PM
from the ask-hn dept.
What I mean by this, will Open Source own the creating of amazing models and cause companies like OpenAI and Anthropic to focus on verticals?I know there's a high costs right now for training models, but I'm sure innovation will change this long term.
I'm genuinely curious about this topic and love learning. My fundamental belief is that it makes sense to build highly tuned models that specialized in specific areas and using orchestration on top to call out to these specialized models with a "shared memory". This would allow for the ability to improve parts of the system without the fear of regression in other areas and it would be cheaper to create a smart "platform". I know there's models with MoE etc but feel like the "microservices" approach (micromodels) would work well here.
I could be totally wrong.
Posted by
mrburton on Monday October 26, 2026 @11:46AM
from the ask-hn dept.
If you're doing fine tuning, how are you curating or creating your training data? If you're doing distillation what tools? Share you're experience.That includes you PewDiePie; Felix Kjellberg.
Posted by
mrburton on Thursday October 22, 2026 @07:39AM
from the reddit.com dept.
If my grandmother had wheels, she would have been a bike.
I'm glade you get stiff reading peoples comments, kudo's to you gangster. Now Daddy Chill.
Who said you can't? What I'm saying is people who look at suicide clearly need help with perspective and support.
If people who feel like terminating their life, got help, I strongly believe there was be a large drop in suicide.
You'll have elaborate on "life is just too much". That to me, opinion here, sounds like "I don't want to try". That's suitable for games, but I think life is something special.
Likewise and this day an age you don't need much to do that.
You really don't need to engage in the thread if it bothers you that much. Do keep in mind, maybe you're on the left side of the Dunning-Kruger effect.
Memento et mori.
Don't rush it.
By the way, 2026 was the 30 year anniversary of his passing. I'm 47 years old and I still think of him almost daily.
To all those reading this thread, suicide isn't the answer. I know you've heard this and it's not special coming from me.
We all heard it's a permanent solution to a temporary problem and so many other quotes. I've lost a lot of people growing up to suicide and the impact it has on those who continue to live on that you love is traumatic.
The man who raised me like I was his son told my mother that he would never commit suicide. He had stage 4 cancer that eventually took his life on my 17th birthday. His reason for enduring through the pain was to teach me that suicide is never the answer because it hurts all those around you.
Although he passed away on my 17th birthday of his illness, he forever changed my life in so many ways. Each and every day I aim to be a better man because of how he loved me.
Please take a moment to utilize this time you have to change those around you. You don't need to create a massive company, be a CEO, have a career people admire, etc.. all you need to do is be someone that inspires others to improve like John did for me.
Posted by
mrburton on Tuesday October 13, 2026 @09:47PM
from the ask-hn dept.
What do you think will be the hot trend in 2027 with on-device AI?
Posted by
mrburton on Thursday July 02, 2026 @12:55PM
from the ask-hn dept.
1. Coding
2. TTS (Text to Speech) and STT (Speech To Text)
3. Image creation and understandingFor me I've been using:1. qwen3-coder-next
2. Fish Audio S2 Pro (TTS) and Whisper for STT
3. Gemma for image analysis and Flux for creation.I run these on my 16-inch MacBook Pro (Apple M5 Max chip with 18-core CPU, 40-core GPU, and 16-core Neural Engine
) with 128GB unified memory
I can relate to this and in my journey, I was able to maintain feeding my need to be useful to the world around me and do well money-wise. The latter was my desire to be there for my family so they can focus on their health.
The best part about HackerNews, is that you get a very good sense of the envious and jealous nature of people. A lot of the "hate" or "angry" comments, are basically people who hate their relationship with "work".
To the author, I think you'll continue that process of being useful, but you'll see that in this new world, you're usefulness now scales..
R.I.P Joe Armstrong - he was an amazing man, engineer, and human overall. :(
Posted by
mrburton on Wednesday September 24, 2025 @03:24PM
from the self-post dept.
A quick write up since I see so many people who don't know the difference. Now that AI is writing a lot more code, I wanted to share something useful for humans.In HTML, div and span are fundamental elements but differ in their roles due to display behavior and use cases. A div is a block-level container (display: block) that takes up the full width of its parent and starts on a new line,...
Posted by
mrburton on Saturday July 19, 2025 @12:58PM
from the ask-hn dept.
How are you preparing data, what challenges do you face, how do you generate data at scale (using LLMs to create synthetic data for instance).Curious minds what to know :)
I don't recall when I first met Bob; I've known him for over 14 years. I've known him first virtually on IRC and then in person at around 2010. We share so many common friends, it's pretty crazy. I guess that's the open source world of nerds for you.
One thing I loved about Bob, is regardless of what he built, who he knew or worked with, you would never see him lean on that to express who he was..
Bob was just a real person with a passion for programming and helping people.
Looking back on my time knowing Bob, I do regret not taking him up on his offer to work with him. I could only imagine how I would have grown as a person by working closely with him.
I do want to say, Bob clearly had a consistency in his style of clothing. It didn't changed over the years.. lol
It all still seems unreal.
Posted by
mrburton on Wednesday January 19, 2022 @06:36PM
from the harness.io dept.
I personally don't have children and I didn't _really_ understand how difficult it is to have kids at home while working. When I moved back East, I did spend 3 days at my brothers.
He has two adorable little girls; 1.5 years old and 2 months.
His kids are pretty calm, by most standards.. but you can't let the older one out of sight for a few seconds.
I certainly feel your pain even though I don't have kids.
Question: Tell me how much you love going to the bathroom or getting washed now. It's the only time you have to yourself lol
Item #1 - 100% correct.
Item #4 - I fully disagree. Some choices are scary to actually make, but are healthy. For instance, my mom stayed with her abusive husband for many years because she had Stockholm syndrome. If you had asked her why didn't she leave during that time, she would say "I didn't have a choice". Ask her now, and she will tell you "Fear - fear of not knowing how things would work out."
Sometimes the hardest choices to make are the impossible choices.
I was in Austin last year and was shocked by the homeless there. I spoke to a few people and they had informed me that some laws changed recently which allows them to live under bridges and the lots of churches etc.
I was taken back because I never thought Austin would have such a problem.
You might want to kill me after I say this, certainly people on HN want to kill me just for fun ;), but I also drink coffee. I only drink sugar free Red Bull. What's strange is I can stop for months without any withdraw from caffeine. I often go from drinking Red Bull + Coffee for a few months to just water for multiple months.
Once I finish this case, I think it's time to go back to water and coffee for 6 months. :)
Life choices always have an impact. I'm not judging you, because trust me, expensive areas like New York City (I love living there) are great, but blow during a pandemic since you're unable to enjoy what the actual city has to offer.
When I relocated back east, I made it a point to a) not live in the city since it'll be overpriced and I won't be able to enjoy the surrounding area and b) Price per square footage wouldn't make sense if I'm at home 24/7. So I went across the river and pay less for _a lot more_.
I know a lot of people are moving to Austin TX. They get a LOT more for their money. I was lucky to be on a month by month lease when the pandemic hit.
I agree - dressing up was implied with "get washed/dressed". It's very important to get dressed for work.