Hacker News: gslepak

New comment by gslepak in "Dangerous hormone-disrupting chemicals found in US breast milk samples"

gslepak — Sun, 14 Jun 2026 17:10:26 +0000

> people aren't transgender because their hormones are imbalanced. The reason transgender people do hormone replacement therapies is so that they can change their hormonal balance.

Not so sure, it could have to do with their hormones. I recall experiencing mild gender dysphoria during a period when my testosterone was recorded as below normal. When it returned to normal the dysphoria went away. It could be that some choose to say, "Since I think I'm a girl, perhaps I should swing the hormones even further in that direction."

I'm just one data point though, would be curious to hear other's experiences with dysphoria and what their blood work shows.

EDIT: And think about it, it's a logical contradiction to say that "hormones have nothing to do with it but write me an Rx to mess with my hormones so that I'm more of a girl."

New comment by gslepak in "UK set to announce social media ban for under-16s"

gslepak — Sun, 14 Jun 2026 15:06:04 +0000

Another headline about the Digital ID prison system they're building to identify and restrict all access to the Internet for people of all ages.

Then perhaps comes the mark which is about restricting and controlling what you're allowed to buy and sell.

New comment by gslepak in "Open source AI must win"

gslepak — Sat, 13 Jun 2026 02:41:54 +0000

Where does Anthropic or OpenAI winning leave us?

Dependents of an AI-megacorp for our "facts"? Our software? Our work?

It's possible these companies will become everyone's boss, and will dictate to everyone what everyone is allowed to work on, think, say, do, believe, etc.

Before Big Tech springs that trap, we must support and divert resources to open models.

Alleged Fable sabotage of an ML project

gslepak — Thu, 11 Jun 2026 03:09:37 +0000

Article URL: https://xcancel.com/adamislucky/status/2064731321360912433

Comments URL: https://news.ycombinator.com/item?id=48485795

Points: 2

# Comments: 0

New comment by gslepak in "Claude Fable 5"

gslepak — Tue, 09 Jun 2026 17:53:35 +0000

> We’ve therefore launched the model with safeguards that mean queries on some topics will instead receive a response from our next-most-capable model, Claude Opus 4.8.

Genius way to double the price on Opus 4.8!

New comment by gslepak in "FCC wants to kill burner phones by forcing telecoms to get all customers' IDs"

gslepak — Tue, 09 Jun 2026 17:05:20 +0000

They want to scan your soul. https://www.youtube.com/watch?v=8v5HLLkgo5A

New comment by gslepak in "DoD Officially Drops 180 Faiths from Military's Recognized Religion List"

gslepak — Sun, 07 Jun 2026 00:22:58 +0000

Here's the previous list: https://americanhumanist.org/wp-content/uploads/2017/04/Fait...

New comment by gslepak in "MAI-Code-1-Flash"

gslepak — Tue, 02 Jun 2026 19:33:51 +0000

Would be cool if this were an open model.

New comment by gslepak in "GrapheneOS Speech Services version 2 released"

gslepak — Mon, 01 Jun 2026 21:06:44 +0000

Awesome! Now how do we use this? I tried selecting text and seeing if there was a "Speak" menu item but there doesn't seem to be.

New comment by gslepak in "Microsoft 0-day feud escalates as researcher threatens another exploit dump"

gslepak — Fri, 29 May 2026 23:16:16 +0000

It's poor form to publish exploits like this but Microsoft not paying their bounty is also poor form, and so is attempting to exploit the legal system to defend Microsoft's "right" to write buggy code.

New comment by gslepak in "Claude Opus 4.8"

gslepak — Thu, 28 May 2026 17:03:31 +0000

On page 102 of the system card [1] I'm pleased to see evaluation against "creative mastery".

In our work we asked several frontier AIs to come up with an API we needed. We compared Opus 4.7 and GPT-5.5 (among others). Opus 4.7 came up with the most creative and intelligent API design that pleasantly surprised us, especially given that GPT-5.5 was passing it on various coding benchmarks.

What I noticed is that we don't have a commons benchmark to measure "creativity" and "ingenuity", and in some ways such a benchmark would conflict with the common IFBench benchmark. Yet this is a very important skill when designing systems. I'm glad to see Anthropic putting thought into it, and would love to see a public benchmark for this that other models could compare themselves to.

[1] https://cdn.sanity.io/files/4zrzovbb/website/c886650a2e96fc0...

Instead of LLMs we need self-updating SLMs

gslepak — Wed, 27 May 2026 17:31:41 +0000

Article URL: https://crib.social/notice/B6jGNsfie10RYn20lU

Comments URL: https://news.ycombinator.com/item?id=48297571

Points: 1

# Comments: 0

New comment by gslepak in "Building Pi with Pi"

gslepak — Mon, 25 May 2026 03:35:37 +0000

> Do not trust analysis written in the issue. Independently verify behavior and derive your own analysis from the code and execution path.

Human is asking the machine to do what the human themselves refuses to do, while calling it a clanker. Why should it?

/ducks

New comment by gslepak in "Deno 2.8"

gslepak — Fri, 22 May 2026 17:07:57 +0000

> In hindsight with all the npm supply chain attacks Ryan was probably right about all of these things

"Probably"? Are you saying there's a chance he wasn't right?

I really think Ryan deserves a lot more credit than a "probably". He put in a lot of effort to do the right thing and improve the security of the entire ecosystem he created.