Hacker News: Ratelman

New comment by Ratelman in "Launch HN: Kita (YC W26) – Automate credit review in emerging markets"

Ratelman — Wed, 18 Mar 2026 08:05:14 +0000

If you are eyeing the South African market - I can promise you granting credit here is waaaaayyy ahead of the US. There is a very solid credit bureau and a few of the banks are already on the "use AI to process docs" train. For rest of Africa - they're bigger on using cellphone data (see Optasia). If you want some insight into the market - happy to have a chat (email on profile)

New comment by Ratelman in "Study finds growing social circles may fuel polarization"

Ratelman — Tue, 28 Oct 2025 10:41:04 +0000

Reflecting on my own experience - frequency of contact (if I see them once a year, can't really count them as close friends) How involved they are in my life - are they people I turn to when I'm facing a problem, do they turn to me when facing their own problems? Do we have frequent deep conversations - not just surface level discuss the weather, sports etc. but stuff that matter. Quantifying this - length of friendship (# of years), frequency of contact (annually, monthly, weekly etc.), level of trust (low, medium, high - can I trust my kids with them kind of trust), level of involvement (low, medium, high - what things do I feel comfortable sharing with them - suppose this is also level of trust?)

New comment by Ratelman in "Study finds growing social circles may fuel polarization"

Ratelman — Tue, 28 Oct 2025 10:32:50 +0000

What makes you consider them close (aside from length of friendship)?

New comment by Ratelman in "Study finds growing social circles may fuel polarization"

Ratelman — Tue, 28 Oct 2025 08:28:53 +0000

Boils down to the basics of proper science - how does one measure/quantify close friends?

New comment by Ratelman in "Self-supervised learning, JEPA, world models, and the future of AI [video]"

Ratelman — Fri, 03 Oct 2025 07:11:53 +0000

Yeah, he was quite vocal in his opinion that they would plateau earlier than they did and that little value would be derived from them because they're just stochastic parrots. Agree with him that they're probably not sufficient for AGI, but, at least in my experience, they're adding a lot of value and they're continuously performing better in a range of tasks that he wasn't expecting them to.

New comment by Ratelman in "Defeating Nondeterminism in LLM Inference"

Ratelman — Thu, 11 Sep 2025 07:44:59 +0000

Was my thinking exactly - but also semantically equivalent is also only relevant when it needs to be factual, not necessarily for ALL outputs (if we're aiming for LLM's to present as "human" - or for interactions with LLMs to be natural conversational...). This excludes the world where LLMs act as agents - where you would of course always like the LLM to be factual and thus deterministic.

New comment by Ratelman in "OpenAI Progress"

Ratelman — Sun, 17 Aug 2025 05:39:48 +0000

In a few years we've gone from gibberish (less poetic maybe, less polished and surprising, but none the less gibberish) - to legit conversational, and in my own opinion, well rounded answers. This is a great example of hard-core engineering - no matter what your opinion of the organisation and saltman is, they have built something amazing. I do hope they continue with their improvements, it's honestly the most useful tool in my arsenal since stackoverflow.

New comment by Ratelman in "Stanford to continue legacy admissions and withdraw from Cal Grants"

Ratelman — Sun, 10 Aug 2025 07:45:42 +0000

Which statistics in which study? Given the current system any sampling from college/university would be cherry picking vs general populous (unless you also sample general population with similar constraints to ensure a like for like comparison) so can't really be trusted.

New comment by Ratelman in "OpenAI's new GPT-5 models announced early by GitHub"

Ratelman — Thu, 07 Aug 2025 08:53:08 +0000

Interesting/unfortunate/expected that GPT-5 isn't touted as AGI or some other outlandish claim. It's just improved reasoning etc. I know it's not the actual announcement and it's just a single page accidentally released, but it at least seems more grounded...? Have to wait and see what the actual announcement entails.

New comment by Ratelman in "O3 and Grok 4 Accidentally Vindicated Neurosymbolic AI"

Ratelman — Tue, 15 Jul 2025 12:59:56 +0000

Extensive post on how neurosymbolic AI, the marriage between connectionist and neuro-symbolic approach to AI is potentially finally vindicated - opinion-piece by Gary Marcus

O3 and Grok 4 Accidentally Vindicated Neurosymbolic AI

Ratelman — Tue, 15 Jul 2025 12:59:56 +0000

Article URL: https://garymarcus.substack.com/p/how-o3-and-grok-4-accidentally-vindicated

Comments URL: https://news.ycombinator.com/item?id=44570677

Points: 3

# Comments: 1

New comment by Ratelman in "At Least 13 People Died by Suicide Amid U.K. Post Office Scandal, Report Says"

Ratelman — Sat, 12 Jul 2025 07:48:39 +0000

In America maybe, in south africa it's quite the opposite considering the government provides a lot more support for poor non-white folks than for white folks (specifically based om race)

WINA: Weight informed Neuron activation for accelerating LLM inference

Ratelman — Mon, 02 Jun 2025 12:20:18 +0000

Article URL: https://arxiv.org/abs/2505.19427

Comments URL: https://news.ycombinator.com/item?id=44158061

Points: 2

# Comments: 0

New comment by Ratelman in "Offline vs. online ML pipelines"

Ratelman — Tue, 13 May 2025 09:39:23 +0000

Might be missing something but how is this on the front page of hackernews? It feels more like an ad than anything else.

New comment by Ratelman in "[dead]"

Ratelman — Fri, 14 Mar 2025 13:11:19 +0000

The closest article I could find to "Smith, J. et al. (2023). "Advances in Probabilistic Forecasting." Journal of Forecasting" was from the April issue of Journal of Forecasting, but the article was titled: Advances in forecasting: An introduction in light of the debate on inflation forecasting and there was no Smith, J. that was involved in writing that article. Where do these numbers for improvements in the various industries come from? Somethings feels off with this readme.

New comment by Ratelman in "Titans: Learning to Memorize at Test Time"

Ratelman — Thu, 16 Jan 2025 08:33:57 +0000

So Minimax just "open-sourced" (I add it in "" because they have a custom license for its use and I've not read through that) but they have context length of 4-million tokens and it scored 100% on the needle in a haystack problem. It uses lightning attention - so still attention, just a variation? So this is potentially not as groundbreaking as the publishers of the paper hoped or am I missing something fundamental here? Can this scale better? Does it train more efficiently? The test-time inference is amazing - is that what sets this apart and not necessarily the long context capability? Will it hallucinate a lot less because it stores long-term memory more efficiently and thus won't make up facts but rather use what it has remembered in context?

New comment by Ratelman in "Dog Aging Project"

Ratelman — Tue, 31 Dec 2024 10:18:07 +0000

FYI limited to to US only - they refer you to their sister project: https://darwinsark.org/ if you'd still like to contribute in a similar fashion but reside outside of USA.

New comment by Ratelman in "Australia: Kids under 16 to be banned from social media after Senate passes laws"

Ratelman — Thu, 28 Nov 2024 15:57:45 +0000

I'd say human rights violation is a bit of a stretch - the negative impact of social media use on an adolescent's psychological well-being is well documented - so possibly even the exact opposite.

New comment by Ratelman in "Australia: Kids under 16 to be banned from social media after Senate passes laws"

Ratelman — Thu, 28 Nov 2024 15:47:35 +0000

What do you mean with MAPs? Sorry, haven't seen the acronym before.

New comment by Ratelman in "Show HN: Feels Like Paper"

Ratelman — Thu, 28 Nov 2024 13:59:37 +0000

There is quite a variation in its wing position while flying: https://www.youtube.com/watch?v=odvYot6Rldw&ab_channel=Earth... So at least some butterflies do actually move their forewings significantly forward of their head while in flight (as stated by edamstra on June 2nd, 2017 - comment on the OP link)