Hacker News: whoistraitor

New comment by whoistraitor in "A Map of British Dialects (2023)"

whoistraitor — Sat, 19 Apr 2025 09:30:43 +0000

Yeh it’s strange it includes cockney so prominently. It isn’t really very present unless you spend time around the various gentlemen frequenting sports pubs and pie and mash shops in east London, or if you take a black cab very often. I’d say the “roadman” dialect, mixing cockney and Jamaican patois, plus grime vibes, is FAR more common. I’ll hear it everyday wandering around South and east London. I guess it’s a London dialect so it’s in that umbrella,… but how come cockney gets such a fat slab of land?

New comment by whoistraitor in "The Illustrated DeepSeek-R1"

whoistraitor — Mon, 27 Jan 2025 22:43:00 +0000

It’s remarkable we’ve hit a threshold where so much can be done with synthetic data. The reasoning race seems an utterly solvable problem now (thanks mostly to the verifiability of results). I guess the challenge then becomes non-reasoning domains, where qualitative and truly creative results are desired.

New comment by whoistraitor in "Supreme Court rules to uphold TikTok ban, setting stage for shutdown"

whoistraitor — Fri, 17 Jan 2025 16:57:43 +0000

IMO first-or-not is moot. It’s estimated that around one billion people speak English to a reasonably fluent level. Included in that is many of the commonwealth countries in which English often holds second spot as a lingua franca (eg. India). It’s an incredibly global language.

New comment by whoistraitor in "llms.txt directory"

whoistraitor — Mon, 23 Dec 2024 20:22:31 +0000

Perplexity is listed, but do they actually abide by llms.txt? And how can we prove they do? Is it all good faith? I wish there were a better way.

New comment by whoistraitor in "OpenAI O3 breakthrough high score on ARC-AGI-PUB"

whoistraitor — Fri, 20 Dec 2024 18:37:42 +0000

The general message here seems to be that inference-time brute-forcing works as long as you have a good search and evaluation strategy. We’ve seemingly hit a ceiling on the base LLM forward-pass capability so any further wins are going to be in how we juggle multiple inferences to solve the problem space. It feels like a scripting problem now. Which is cool! A fun space for hacker-engineers. Also:

> My mental model for LLMs is that they work as a repository of vector programs. When prompted, they will fetch the program that your prompt maps to and "execute" it on the input at hand. LLMs are a way to store and operationalize millions of useful mini-programs via passive exposure to human-generated content.

I found this such an intriguing way of thinking about it.

New comment by whoistraitor in "Ask HN: SWEs how do you future-proof your career in light of LLMs?"

whoistraitor — Tue, 17 Dec 2024 11:17:09 +0000

Curious: what kind of problem domain to you work on? I use LLMs every day on pretty hard problems and they are always net positive. But I imagine they’re not well trained on material related to your work if you don’t find them useful.

New comment by whoistraitor in "Review of Mullvad VPN"

whoistraitor — Thu, 12 Dec 2024 00:34:08 +0000

It is probably the most reliable yeh, tho spending time here I’ve grown increasingly aware that the great firewall is more than aware of this vpn traffic, even if it’s wrapped up to look like normal traffic. They periodically will seem to ‘dial down’ the internet, especially at politically sensitive times. They are fully aware great swathes of the populace and visitors use VPNs, and they choose to allow it. They’d rather control and monitor than inspire even more opaque channels.

New comment by whoistraitor in "ZombAIs: From Prompt Injection to C2 with Claude Computer Use"

whoistraitor — Sun, 27 Oct 2024 05:25:33 +0000

Don’t know what OP might suggest but my first take is: never allow unstructured output from one LLM (or random human) of N privilege as input to another of >N privilege. Eg, use typed tool/function calling abstractions or similar to mediate all interactions to levers of higher privilege.

New comment by whoistraitor in "Zamba2-7B"

whoistraitor — Tue, 15 Oct 2024 00:35:18 +0000

Ah apologies I misread the architecture. But it does fit the spirit of finding disproportionately higher performance in smaller networks. Still promises of finding smaller sub networks. Running on mediocre mobile devices doesn’t seem a dream when stuff like this is released. Exciting!

New comment by whoistraitor in "Zamba2-7B"

whoistraitor — Tue, 15 Oct 2024 00:23:54 +0000

Cool! Seems we’re moving closer and closer to realizing the Lottery Ticket Hypothesis https://arxiv.org/abs/1803.03635

New comment by whoistraitor in "The New Internet"

whoistraitor — Sat, 27 Jul 2024 07:24:05 +0000

Ie. A form of perverse incentive or the cobra effect. Endemic to capitalism, especially in infrastructure.

New comment by whoistraitor in "Multi-agent chatbot murder mystery"

whoistraitor — Wed, 10 Jul 2024 17:30:39 +0000

You can use HuggingFace, PerplexityAI, Anyscale, TogetherAI. There are a bunch of options. Many offer OpenAI compatible APIs too.

New comment by whoistraitor in "Taking a closer look at AI's supposed energy apocalypse"

whoistraitor — Fri, 28 Jun 2024 18:58:39 +0000

What I’d love to see is a rundown of new efficiencies in generative AI. With things like quantization and specialized transformer hardware, the costs will hopefully be less shocking in the future. And fwiw, I don’t find the article’s stats very shocking anyway. At least it’s a net win type of game, whereas the previous environmental bugbear, PoW crypto, was entirely premised on net-energy loss.

New comment by whoistraitor in "How do our brains adapt to control an extra body part?"

whoistraitor — Sun, 02 Jun 2024 06:02:26 +0000

I would say actually in the vast majority of cases of accessibility interventions, it benefits everyone. Everyone is, at various points in their lives, a varying degree of disabled (defined in terms of functional deficits). Whether injured, pregnant, chronically ill, a wheelchair user, elderly, or — heh — even intoxicated, thinking broadly about users of the spaces and tools we design is always going to yield more positive than negative externalities imho.

In cases where this is not true, I’d challenge designers and engineers to find more novel designs that can genuinely be used by anyone and not cost too much inconvenience. This can sometimes mean starting from scratch and questioning our assumptions. Eg. Even the need for a curb-to-street indicator presupposes a street that is used by both big metal vehicles and pedestrians, whereas perhaps there’s a solution that means those paths never cross. Ie. More fundamental urban and transport design instead of band-aiding atop legacy systems.

New comment by whoistraitor in "OpenAI departures: Why can’t former employees talk?"

whoistraitor — Sat, 18 May 2024 07:28:00 +0000

Indeed. I’ve heard first hand accounts that would make it impossible for me to trust him. He’s very good at the game. But I’d not want to touch him with a barge pole.