Hacker News: youngprogrammer

New comment by youngprogrammer in "The Fall of Scientific American"

youngprogrammer — Mon, 26 Oct 2026 06:14:03 +0000

Atlantic and New Yorker my two favorite publications for long form articles. Although the Atlantic has got a bit too political for my tastes

New comment by youngprogrammer in "I used sound waves to make espresso. It could cut coffee‑brewing energy use by ¾"

youngprogrammer — Fri, 19 Jun 2026 18:46:50 +0000

At a 3 minute shot, I’d rather use the same time to do a pour over

New comment by youngprogrammer in "My new obsession: A horse-racing board game of pure luck"

youngprogrammer — Thu, 28 May 2026 08:17:21 +0000

You might as well play "who can memorize the most openings and lines"

New comment by youngprogrammer in "Cloudflare Flagship"

youngprogrammer — Wed, 27 May 2026 01:11:01 +0000

It can get complicated quickly if you're actually using it in a production system. At my prev enterprise saas company we had feature flags that could be turned on per customer / per environment (dev, staging, prod) with permission + logging model such that our support team could also toggle flags with history of who turned on what. We also had "per user" feature flags for certain test users at companies and had DSL rules to evaluate the features

New comment by youngprogrammer in "Following the Text Gradient at Scale"

youngprogrammer — Wed, 06 May 2026 23:54:22 +0000

this is essentially how bindcraft works for drug discovery: https://www.nature.com/articles/s41586-025-09429-6 (minus the accumulation step)

feedback from af2 folding confidence + structural scoring

New comment by youngprogrammer in "A Brief History of Fish Sauce"

youngprogrammer — Mon, 20 Apr 2026 04:07:49 +0000

Fish sauce is delicious but had to stop using it since it's high in histamine (gives me a stuffy nose) and potentially carcinogenic due to its high levels of nitrosamines

New comment by youngprogrammer in "Scaling Karpathy's Autoresearch: What Happens When the Agent Gets a GPU Cluster"

youngprogrammer — Thu, 19 Mar 2026 23:58:02 +0000

industrialized overfitting is basically what ML researchers do

New comment by youngprogrammer in "Show HN: AI Timeline – 171 LLMs from Transformer (2017) to GPT-5.3 (2026)"

youngprogrammer — Tue, 24 Feb 2026 02:03:12 +0000

should go a bit earlier with word2vec, NMT, seq2seq, attention, self attention

Scientists Contact Whales in World-First Communication Experiment

youngprogrammer — Tue, 19 Dec 2023 17:38:48 +0000

Article URL: https://www.businessinsider.com/how-humans-could-communicate-with-extraterrestrial-aliens-practicing-on-whales-2023-12

Comments URL: https://news.ycombinator.com/item?id=38698690

Points: 1

# Comments: 0

New comment by youngprogrammer in "Ask HN: AI/ML papers to catch up with current state of AI?"

youngprogrammer — Sat, 16 Dec 2023 01:16:18 +0000

Little late to this thread but from my list:

LLM (foundational papers)

* Attention is all you need - transformers + self attention

* BERT - first masked LM using transformers + self attention

* GPT3 - big LLM decoder (Basis of gpt4 and most LLM)

* Instruct GPT or TKInstruct (instruction tuning enables improved zero shot learning)

* Chain of Thought (improve performance via prompting)

some other papers which are become trendy depending on your interest

* RLHF - RL using human feedback

* Lora - make models smaller

* MoE - kind of ensembling

* self instruct - self label data

* constitutional ai - self alignment

* tree of thought - like CoT but a tree

* FastAttention,Longformer - optimized attention mechanisms

* React - agents

New comment by youngprogrammer in "Bad characters: imperceptible Natural Language Processing attacks [pdf]"

youngprogrammer — Thu, 30 Dec 2021 09:47:28 +0000

It seems like most of these imperceptible changes could be addressed by something like ascii folding (https://www.elastic.co/guide/en/elasticsearch/reference/curr...) but this might not apply for non-english use cases.

If you're interested in adversarial NLP, I also recommend reading this blog post on adversarial attacks on GPT2 with universal triggers (e.g. adding "nobody" as prefix for all inputs causes all entailments to be predicted as contradiction).

New comment by youngprogrammer in "Asana S-1"

youngprogrammer — Tue, 25 Aug 2020 01:15:04 +0000

They probably didn't have a Gantt chart to help them figure out the dependencies to properly plan it on their roadmap

New comment by youngprogrammer in "Artificial intelligence is quietly disrupting the fragrance development process"

youngprogrammer — Mon, 06 Apr 2020 06:38:19 +0000

You could do something similar to how they trained a ML model to find antibiotics compounds: https://www.cell.com/action/showPdf?pii=S0092-8674%2820%2930.... First, train a deep learning model to learn a representation of molecules from their molecule structures. Then feed in the thousand or so known compounds that produce pleasant or unpleasant smells as training data with some score of "pleasantness". We can then use this model to quickly score millions of compounds and select candidates to test.

New comment by youngprogrammer in "Coronavirus has disrupted supply chains for nearly 75% of U.S. companies"

youngprogrammer — Wed, 11 Mar 2020 17:55:51 +0000

Anecdote: Our caterer in silicon valley said there was a supply issue for tofu.

Video of Combat Robots Flagged as Animal Abuse

youngprogrammer — Tue, 20 Aug 2019 17:27:14 +0000

Article URL: https://support.google.com/youtube/thread/12332488?hl=en

Comments URL: https://news.ycombinator.com/item?id=20749542

Points: 1

# Comments: 0

Show HN: Num2vec – Numerical Embeddings via Deep Nested RNNs

youngprogrammer — Thu, 01 Feb 2018 11:37:51 +0000

Article URL: http://blog.ayoungprogrammer.com/2018/01/deep-recurrent-neural-networks-for-mathematical-sequence-prediction.html/

Comments URL: https://news.ycombinator.com/item?id=16281112

Points: 6

# Comments: 0

New comment by youngprogrammer in "Stocks with Outperform Ratings Beat the Market"

youngprogrammer — Sun, 13 Aug 2017 15:31:58 +0000

This is nothing like the sharpshooter fallacy. The analysis determined the average performance of the analysts with >100 stocks rated and 10 analysts out of 16 did better than the rest.

New comment by youngprogrammer in "Stocks with Outperform Ratings Beat the Market"

youngprogrammer — Sun, 13 Aug 2017 07:03:18 +0000

How do you measure top ten if you don't have a ranking system?

New comment by youngprogrammer in "Stocks with Outperform Ratings Beat the Market"

youngprogrammer — Sun, 13 Aug 2017 07:02:40 +0000

Top 10 performers out of 16 or so analysts in the analysis is not survivor bias.

New comment by youngprogrammer in "Stocks with Outperform Ratings Beat the Market"

youngprogrammer — Sun, 13 Aug 2017 07:02:03 +0000

Outliers were removed to get a better measure of the "accuracy" of the price targets.

10 day windows were used to reduce the amount of volatility/noise in a time frame

Return horizons for 1 years was used because price targets are for one year.

Theres only 15 or so analysts I looked at.

I was doing this as an exploratory data analysis and didn't want to pull out my old stats textbook.

Cutoffs were chosen to reduce volatility of measurements since I was looking at percentages. A stock going from $1.5 to $2.0 is a 33% increase whereas the movement of $100 to $133 is significantly more impactful. Stock with lower market cap have more volatility. The minimum analyst rating was chosen to eliminate analysts with very small number of ratings as they would be unreliable.