lamroger

8 ms·

61.
by lamroger 3y ago
Cellular automata bridged math and CS for me and it's beautifully chaotic. It's the same backwards with deep learning. Chaos to structure in classifying numbers or "doing" math. It's the unexpected that's excit
62.
by lamroger 3y ago
The guardrails are cool! I think more details of where the data goes and when it goes from few-shot to fine-tune will be helpful.
63.
Ask HN: What's the Trend on Offshoring?
1 points by lamroger 3y ago|0 comments
64.
by lamroger 3y ago
Philly replaced a burnt bridge pretty quickly. Not sure how different the sizes are but hoping for a speedy fix. https://www.businessinsider.com/philadelphia-turns-heads-spe...
65.
by lamroger 3y ago
I wish we had some insight on offshoring numbers
66.
by lamroger 3y ago
The general consensus imo is that fine-tuning is more for tone and style vs accuracy. People use vector DBs to grab relevant data to throw into the prompt and call it Retrieval Augmented Generation. From what this seems to do is host multip
67.
by lamroger 3y ago
Wanted to hack today but the universe is telling me to go enjoy the sun
68.
by lamroger 3y ago
Reminder that vets have it extra hard. And a reminder to myself to do good.
69.
Show HN: Script connecting Bluetooth thermometer to Artisan roasting software(github.com)
1 points by lamroger 3y ago|0 comments
70.
by lamroger 3y ago
really enjoyed this too
71.
by lamroger 3y ago
Wild
72.
Show HN: LLM RAG-style coffee recommender(tinywaffle.co)
1 points by lamroger 3y ago|0 comments
73.
by lamroger 3y ago
Thanks for sharing - incredibly inspirational on a personal and professional level
74.
by lamroger 3y ago
I'd order pizza from that site
75.
by lamroger 3y ago
Sorry - don't have actual advice. I remember the developer relations community being inclusive. Maybe you can DM some of them for specific advice or just someone to talk to. https://twitter.com/mipsytipsy comes to mind
76.
by lamroger 3y ago
it's sad how many places are only staffing with senior+ engineers.
77.
by lamroger 3y ago
No thanks
78.
by lamroger 3y ago
Looks fun!
79.
by lamroger 3y ago
I thought llama is open weights and open to commercial with reasonable restrictions. I haven't heard practitioners who fine tune llama having trouble with it
80.
by lamroger 3y ago
interesting though. when will these robots be walking ads like instacart pivoted to. side hustle selling data to ai companies. it's all from public spaces so no one to argue copyright. i'm in LA and the robots are pretty freaking
81.
by lamroger 3y ago
data science is more = to statistics
82.
by lamroger 3y ago
I feel like modern ML is finding out that you can apply the chain rule to calculate derivatives cost effectively and then scaling up a really really really big equation to then iteratively converge on the answer you want.
83.
by lamroger 3y ago
Agreed! I think it's going to be hard for software developers to adjust to the "data science/engineering" mindset of monitoring and iterating on the long tail of maintenance. A lot of teams already have this issue with d
84.
by lamroger 3y ago
Still incredible in 2023. Goes to show UX and maybe first to market will be the differentiator
85.
by lamroger 3y ago
What is it called when people label themselves as something but aren't that thing? People gotta walk the walk
86.
by lamroger 3y ago
I am - Great for solo dev-ing.
87.
by lamroger 3y ago
thanks for the rec
88.
by lamroger 3y ago
low 7 figures after how long?
89.
by lamroger 3y ago
for sure but going from 90% to 95% might be worth for some
90.
by lamroger 3y ago
``` When to use fine-tuning: Fine-tuning GPT models can make them better for specific applications, but it requires a careful investment of time and effort. We recommend first attempting to get good results with prompt engineering, prompt c

More ›