Hacker News: gbnwl

New comment by gbnwl in "Pacing model development in an era of cyber-critical capabilities"

gbnwl — Thu, 20 Aug 2026 04:56:25 +0000

It seems like the entire thought process you’re trying to sell hinges on the idea that OpenAI reported the attack first. Did you forget that it was actually HuggingFace that reported it first, and OpenAI only stepped forward latter?

New comment by gbnwl in "SQLite Critical CVEs or LLM Slop?"

gbnwl — Mon, 03 Aug 2026 18:24:44 +0000

It’s just a boring and unhelpful complaint that afaict largely serves to soothe the commenters ego rather than point at anything insightful that’s useful or predictive. Point me to your favorite “next token predictor” comment that was actually insightful or predictive. You have years of material to draw from.

New comment by gbnwl in "SQLite Critical CVEs or LLM Slop?"

gbnwl — Mon, 03 Aug 2026 17:40:32 +0000

I just prefer HN comments to be better reflections of reality. There is an unspoken expectation here that people here know what they’re talking about especially when it comes to technical matters. The rise of LLMs has given way to a HN branded populism that willingly denies reality as well. “Next token predictor” truthism is just so dumb and completely ignores the reality of what these tool are able to do. Smash that upvote button every time it feels good if you want but it’s just a meaningless take at this point. It won’t help you predict anything that’s coming.

Since we disagree on the present let’s informally do a “remind me 2 years” to this discussion and see what’s happened then.

New comment by gbnwl in "SQLite Critical CVEs or LLM Slop?"

gbnwl — Mon, 03 Aug 2026 17:17:51 +0000

Fair so let me be clear. I’m whining because the “next token predictor” reductionist point of view has been wrong and is only growing more wrong with time. Clearly these things can do things that actually matter. Do you disagree?

New comment by gbnwl in "SQLite Critical CVEs or LLM Slop?"

gbnwl — Mon, 03 Aug 2026 16:47:57 +0000

Are they useful or not? Will they continue changing the world or not? People who choose one way or the other for describing them typically fall on one side or the other in these questions imo. What do you think? Will these next token predictors change the world or not?

New comment by gbnwl in "SQLite Critical CVEs or LLM Slop?"

gbnwl — Mon, 03 Aug 2026 16:26:32 +0000

Every day I wake up and open HN.

“LLM has made legitimate mathematical discoveries” —> Wow the rate of progress is amazing. Highly upvoted.

“LLM does something not good” -> Does everyone else not realize LLMs are just dumb next token predictors? Highly upvoted.

So tired of this discourse and this site.

New comment by gbnwl in "Ten advances in mathematics and theoretical computer science"

gbnwl — Sat, 01 Aug 2026 16:14:10 +0000

There are articles with far fewer upvotes and comments ranking higher on the front page right now, despite being the same age or older than this one. HNs opaque ranking system at it again.

New comment by gbnwl in "Ten advances in mathematics and theoretical computer science"

gbnwl — Sat, 01 Aug 2026 16:11:32 +0000

Everyday? Which 10 problems were solved by mathematics grad students in the past 10 days?

OK I’ll grant that it’s not your obligation to be my search function (despite you making the wild assertion in the first place), so instead can you just point us to the latest grad student solved problem of this level that you know of?

New comment by gbnwl in "Advancing the price-performance frontier with GPT‑5.6"

gbnwl — Thu, 30 Jul 2026 23:04:38 +0000

I think the fundamental difference between our assumptions is you believe prompts to be optimized for tasks rather than model-task pairs. The only elaboration I can give you is empirical observations and model providers own guidance (as someone has already linked here). I'm pretty sure you probably have specific parts of your prompts that came about due to specific failure modes observed in your evals of running the task against first model. These vary across models in my experience, and it's always worth redoing this calibration process.

New comment by gbnwl in "Advancing the price-performance frontier with GPT‑5.6"

gbnwl — Thu, 30 Jul 2026 18:15:12 +0000

Experienced similar between 5.4-mini vs 5.6-luna in our own pipelines but after spending some time on prompt optimization and testing out various reasoning effort levels 5.6-luna was well worth it. Did you just replace model selection while keeping everything else in place or spend some time on evaling with newer prompts etc?

New comment by gbnwl in "Grok 4.5"

gbnwl — Wed, 08 Jul 2026 19:57:38 +0000

We can assume (outside of Ollama) that they meant the strongest model from each lab. If you limit yourself to just looking at the literal strings in the list, literally none of these are models. What model is "Deepseek" or "GPT"?

New comment by gbnwl in "GPT-5.6 Sol Ultra will be in Codex"

gbnwl — Mon, 06 Jul 2026 04:43:41 +0000

How is this any different than what we have already? We've had this ability for ages (6+ months, decades in the AI world), you can literally today easily prompt CC or Codex to use subagents to accomplish tasks and they'll do it well. My entire workflow is one top level orchestrator chat creating tickets to dispatch to subagents to implement, and other subagents to verify. Why is this being sold as a new thing? Have HN users never tried tried asking CC or Codex to use subagents?

New comment by gbnwl in "Medical students are using popular research tool to pump out misleading studies"

gbnwl — Thu, 25 Jun 2026 10:22:11 +0000

As usual HN posters are hyper aware of other's credentials while ignoring that their BS in CS (if that) doesn't magically qualify them to assess everything in every domain.

"I'm a software engineer, I'm sure if I had the time to study Neuroscience, I'd figure out what all of these researchers failed to realize all these decades! I (alone) have the magic of critical and logical thinking"

New comment by gbnwl in "Codex logging bug may write TBs to local SSDs"

gbnwl — Mon, 22 Jun 2026 13:11:35 +0000

https://xcancel.com/trq212/status/2014051501786931427#m

New comment by gbnwl in "US Scientist John Jumper to Leave Google DeepMind for Anthropic"

gbnwl — Sat, 20 Jun 2026 16:08:09 +0000

Hm first Shazeer and now Jumper, DeepMind getting hollowed out this week.

New comment by gbnwl in "Alphabet announces $80B equity capital raise to expand AI infra and compute"

gbnwl — Mon, 01 Jun 2026 21:23:54 +0000

Don't forget that the denominator (total number of outstanding shares) will be increased by this as well. So even if the market cap reacted exactly one to one like you're proposing the per share price wouldn't stay constant necessarily.

New comment by gbnwl in "Real-time LLM Inference on Standard GPUs: 3k tokens/s per request"

gbnwl — Fri, 29 May 2026 20:01:45 +0000

No they didn't, they predict they'll get that much. Also worth noting the prediction assumes running at MXFP4/FP8 quantization.

New comment by gbnwl in "Frontier AI has broken the open CTF format"

gbnwl — Sat, 16 May 2026 21:06:49 +0000

Frontier as in "Frontier Model" is a legitimate vocabulary term you should probably be aware of in 2026. It's not something the author made up or chose randomly, it's common parlance in the space.

New comment by gbnwl in "The old world of tech is dying and the new cannot be born"

gbnwl — Fri, 15 May 2026 23:29:23 +0000

Can you really look yourself in the mirror and say with a straight face that fundamentally nothing has changed about the relationship between the US and its allies? Do you really think Europeans will be quick to forgive the wrongs of this administration? They’ve lost faith in our political system and will, rightfully so, do everything in their power to disentangle with us. The problem with your theory is that they know even if Trump is replaced by someone closer to European social values, our electorate could just as easily completely reverse course in 4 years. It literally already happened. Bush never threatened to annex European territory with military force as far as I can tell. But I understand why in these chaotic times you’d want to gravitate towards hopeful fictions.

New comment by gbnwl in "The old world of tech is dying and the new cannot be born"

gbnwl — Fri, 15 May 2026 23:23:36 +0000

Probably a better way to phrase it would be keeping “pace“. Yes, they are still behind but by about the same amount as they always have, they aren’t drifting further behind. Like two marathon runners, one a a mile ahead, but both maintaining the same mile time.