HNHacker News
TopNewBestAskShowJobs

d_burfoot

2,282 karma · joined March 20, 2016

Big Data and ML engineer, formerly at LinkedIn, now building WebWidgets.io.

https://www.linkedin.com/in/daniel-burfoot/

Email : first.last@gmail.com

submissionscomments
d_burfoot··on AI optimism is a class privilege
> they mimic and amplify the inherent racism present in their own training data

LLMs turn out to be biased against white men:

https://www.lesswrong.com/posts/me7wFrkEtMbkzXGJt/race-and-g...

> When present, the bias is always against white and male candidates across all tested models and scenarios. This happens even if we remove all text related to diversity.

d_burfoot··on Ask HN: How can we solve the loneliness epidemic?
I mean this seriously: we need more cults.

Cults have been viciously slandered by mainstream information sources, often because lurid cult stories generate clicks and headlines. Of course some cults are abusive, just like some marriages are abusive. But we still think marriage is good in general.

If you think all cults are bad, you're implicitly against all religion, since every mainstream religion was once a cult. Being anti-cult is also profoundly un-American. America was built by cultists. Freedom of religion is literally the first principle stated in the Bill of Rights.

A cult is really just a professionally managed social environment. If you trust professionals like lawyers, doctors, or teachers with their respective duties, there's no reason in principle you shouldn't trust a cult leader to manage your social environment for you. Of course you should vet them, ask about their reputation, etc.

d_burfoot··on Fabrice Bellard's TS Zip (2024)
Kolmogorov Complexity is only defined up to a constant, which represents Turing machine translation length.
d_burfoot··on C Is Best (2025)
Given SQLite's amazing battery of test code, it seems like it would be valuable, illuminating, and not too difficult to write a test implementation in Rust. Can you get the Rust implementation to pass the test suite? Either answer to that question would be interesting.
d_burfoot··on Claude Code On-the-Go
I really want to use Claude Code on the phone or tablet, with voice commands only, and perhaps a few simple approval thumb actions. I don't want to type out complex prompt information on a virtual keyboard. I tried setting this up with some of the iOS terminal emulators, and it almost worked, but there was some glitch where Claude would try to start using the first characters that arrived from the voice command.

Anyone have better results?

d_burfoot··on Inside the proton, the ‘most complicated thing you could possibly imagine’ (2022)
When I was a physics student, there were four forces: strong, weak, EM, and gravity. That picture seemed neat and clean. Strong kept the nucleus together, EM kept molecules and atoms together (or broke them apart), gravity kept astronomical bodies together, weak was some kind of momentum-accounting device.

Recently, GPT informed me that the strong force is really a tiny after-effect of the "QCD force" (in the same way that the Van der Waals forces are after-effect of EM). Also, more and more questions about "dark matter" seem to be building up, suggesting that the standard Newton-Einstein story of gravity is far from the complete picture.

25 years ago it seemed like physics was mostly complete, and the only remaining work was exploring the corner cases and polishing out all the imperfections. It doesn't feel that way anymore! The confusing part is that modern physics is so unbelievably successful and useful for technology - if the underlying theory was way off, how could the tech work?

d_burfoot··on A quarter of US-trained scientists eventually leave
Second sentence of abstract:

> Leave rates are lower in the life sciences and higher in AI and quantum science but overall have been stable for decades

The US has been completely dominant in technology innovation for the last several decades. So, the answer is no: the loss of 1/4 of the STEM scientists is not important.

d_burfoot··on Wolfram Compute Services
Do quant traders use Mathematica? I would guess this would be a great use case for a tool that lots of people love. Pretty language, huge boatload of built-in tools, high powered mathematics, great visualization capabilities. Quant firms should be able to live with the price tag. I assume they have a compiler that can produce fast executables for HFT.
d_burfoot··on A Remarkable Assertion from A16Z
Hypothesis C: failure of human memory. A human read Stephenson's book(s) 20 years ago, remembers that the endings were a bit unsatisfying. The same human also read some other book many years ago, which ends mid-sentence. In that person's mind, the two are conflated.
d_burfoot··on I don't care how well your "AI" works
> AI systems exist to reinforce and strengthen existing structures of power and violence. They are the wet dream of capitalists and fascists.

Persuasion tip: if you write comments like this, you are going to immediately alienate a large portion of your audience who might otherwise agree with you.

d_burfoot··on Disrupting the first reported AI-orchestrated cyber espionage campaign
Wait a minute - the attackers were using the API to ask Claude for ways to run a cybercampaign, and it was only defeated because Anthropic was able to detect the malicious queries? What would have happened if they were using an open-source model running locally? Or a secret model built by the Chinese government?

I just updated by P(Doom) by a significant margin.

d_burfoot··on AWS multiple services outage in us-east-1
I think AWS should use, and provide as an offering to big customers, a Chaos Monkey tool that randomly brings down specific services in specific AZs. Example: DynamoDB is down in us-east-1b. IAM is down in us-west-2a.

Other AWS services should be able to survive this kind of interruption by rerouting requests to other AZs. Big company clients might also want to test against these kinds of scenarios.

d_burfoot··on What Americans die from vs. what the news reports on
A big chunk, perhaps the majority, of the "Accidents" are from cars. Another infographic I observed recently showed that, for children, the risk of death due to traffic accidents was greater than all other risks combined.

People should be raving and screaming for faster rollout of self-driving cars. If self-driving cars were an experimental drug undergoing a clinical trial, they would cancel the trial at this point because it would be unethical to continue denying the drug to the control group.

d_burfoot··on PEP 810 – Explicit lazy imports
Looks good to me. I use a tab-completion trick where the tab-completer tool calls the script I'm about to invoke with special arguments, and the script reflects on itself and responds with possible completions. But because of slow imports, it often takes a while for the completion to respond.

I could, and sometimes do, go through all the imports to figure out which ones are taking a long time to load, but it's a chore.

d_burfoot··on I got the highest score on ARC-AGI again swapping Python for English
To me the reason ARC-AGI puzzles are difficult for LLMs and possible for humans is that they are expressed in a format for which humans have powerful preprocessing capabilities.

Imagine the puzzle layouts were expressed in JSON instead of as a pattern of visual blocks. How many humans could solve them in that case?

d_burfoot··on I replaced Animal Crossing's dialogue with a live LLM by hacking GameCube memory
> LLM-powered NPCs is one thing I'm most excited about in the future of gaming

Incoming new type of health crisis: video game addiction coupled with LLM-induced psychosis. Dudes spending 12 hours a day farming gold in a MMORPG while flirting with their AI girlfriend sidekick

d_burfoot··on Anthropic raises $13B Series F
There's a big issue with a lot of thinking about these valuations, which is that LLM inference does not need the 5-nines of uptime guarantees that cloud datacenters provide. You are going to see small business investors around the world pursue the following model:

- Buy an old warehouse and a bunch of GPUs

- Hire your local tech dude to set up the machines and install some open-source LLMs

- Connect your machines to a routing service that matches customers who want LLM inference with providers

If the service goes down for a day, the owner just loses a day's worth of income, nobody else cares (it's not like customers are going to be screaming at you to find their data). This kind of passive, turn-key business is a dream for many investors. Comparable passive investments like car washes, real estate, laundromats, self-storage, etc are messier.

d_burfoot··on Use One Big Server (2022)
Any stats on HN downtime over the years? I remember one or two outages in the last decade or so, but I would guess the uptime is about 99.99%.
d_burfoot··on Show HN: Base, an SQLite database editor for macOS
I love this model of software development, which I think of as "artisanal", as opposed to open-source or enterprise. A small team, maybe just one person, who invests deep effort into a product that does one thing very well.
d_burfoot··on Privately-Owned Rail Cars
My wife loves the train (hates driving) and so this would be quite interesting to us. But I've heard too many Amtrak horror stories, like the one about how the train broke down about ten miles away from her destination, and they wouldn't let her get off, so she had to sit there for ten hours until they were able to fix it.
d_burfoot··on I tried every todo app and ended up with a .txt file
The problem with productivity apps is that one size does not fit all. Everyone has radically different goals, constraints, interests, and workflows. Many people would benefit from having a "living" app that is personalized to their tastes, and also adapts over time as the characteristics of your life change (e.g. having kids is probably going to change your approach to productivity!)

I recommend finding a framework within which to build your own apps, and then building your own suite of apps that have exactly the features you want (shameless plug, I built a platform that can serve as such a framework : WebWidgets.io. It is basically a way to hook up SQLite databases to JS objects in the browser, which allows wide liberty to build your own simple custom apps).

d_burfoot··on Show HN: I've been building an ERP for manufacturing for the last 3 years
Does the ERP allow you API access, or do you need to do CSV upload/download?
d_burfoot··on Ask HN: What trick of the trade took you too long to learn?
#1 Rule of Investing: you can't beat the market, but you can beat the tax man
d_burfoot··on Study mode
This is the kind of thing that could have been a decent AI startup - hire some education PhDs, make some deals with school systems, etc.

In the old days of desktop computing, a lot of projects were never started because if you got big enough, Microsoft would just implement the feature as part of Windows. In the more recent days of web computing, a lot of projects were never started, for the same reason, except Google or Facebook instead of Microsoft.

Looks like the AI provider companies are going to fill the same nefarious role in the era of AI computing.

d_burfoot··on Python classes aren’t always the best solution
For people who think classes are terrible, can you explain how you would implement the functionality associated with a Pandas DataFrame, without using classes? Did the Pandas developers simply get it wrong when choosing how to build this?

https://pandas.pydata.org/docs/reference/api/pandas.DataFram...

d_burfoot··on All AI models might be the same
For those who don't know about Plato's Theory of Forms, the remarkable empirical fact is that all humans learn roughly the same concepts representing words like "dog", "house", "person", "boat", and so on, even if those people grew up in different places and never had any overlap between their observational experience. You and I may never have observed the same dog, but we still agree on what a "dog" is.

This phenomenon appears to occur for LLM learning as well, though it is less remarkable due to the fact that LLMs likely have significant overlap in their training data.

I believe this is good news for Alignment, because, as Plato pointed out, one of the most important forms is the Form of the Good - a (theoretical) universal human ideal containing notions of justice, virtue, compassion, excellence, etc. If this Form truly exists, and LLMs can learn it, it may be possible to train them to pursue it (or refuse requests that are opposed to it).

d_burfoot··on Ask HN: What Are You Working On? (June 2025)
This is a good idea. I think email is poised to become a more influential user interface tool, due to the advent of LLMs.
d_burfoot··on Object personification in autism: This paper will be sad if you don't read (2018)
I feel a very deep, apparently irrational reluctance to throw away objects I no longer need, especially if those objects are well-crafted. I feel that doing so is disrespectful of the love and effort the object's creator invested in them.
d_burfoot··on See how a dollar would have grown over the past 94 years [pdf]
Your own link, to the BPP Wiki page, says:

> In 2007, the BPP initiated Inflación Verdadera, a project that provided a daily inflation gauge for Argentina, serving as an alternative to the official consumer price index which was manipulated by the government during 2007-2016

d_burfoot··on See how a dollar would have grown over the past 94 years [pdf]
Note that the government has huge incentives to downplay the inflation rate. Over the time period of the graph, the price of gold went up by about 100x; I would consider that a more accurate estimate of inflation than the 18x number implied by the chart. Due to the way exponentials work, you can hide a lot of inflation over 100 years by claiming the annual rate is just half a percent lower than it really is.
← PreviousPage 2 of 14Next →