Hacker News: blochist

New comment by blochist in "Ask HN: How to be SOC2 Type 2 compliant as a solo-entreprenuer?"

blochist — Fri, 15 May 2026 14:03:26 +0000

SOC2 is, at the end of the day, a voluntary compliance standard. HIPAA and FERPA requirements are federal law. Waiving those requirements would not just mean accepting additional liability, but would normally make your customer ineligible to receive federal funds, which are typically a substantial chunk of revenue.

New comment by blochist in "US inflation jumps to 3.8% as energy costs surge from Iran war"

blochist — Tue, 12 May 2026 14:44:07 +0000

This. Energy is up 17.9% and energy commodities (oil, gas, etc.) 29.2%. See the CPI release: https://www.bls.gov/news.release/cpi.nr0.htm.

New comment by blochist in "Rathbun's Operator"

blochist — Wed, 18 Feb 2026 16:33:57 +0000

It's an excellent comment on the attitude behind the question and this is, after all, a comment section not an "answers" section.

New comment by blochist in "Are You YES AI or No AI?"

blochist — Mon, 19 Jan 2026 17:37:44 +0000

Aside from the fact that it kind of obviously is if you "vote" the fact that it says "Not sure yet? That's okay — vote anyway!" is kind of a give-away that this isn't going to produce anything like rigorously useful data on the question (it produces a lot of other data though!)

New comment by blochist in "The email they shouldn't have read"

blochist — Wed, 08 Oct 2025 16:22:35 +0000

Usually "unilateral amendments" are allowed via the contract terms, so it's part of the original contract.

New comment by blochist in "What GPT-OSS leaks about OpenAI's training data"

blochist — Mon, 06 Oct 2025 16:36:00 +0000

Importantly, though, the token is "\\xadder" which looks a bit like an escaped hex code. That actually suggests a different origin of the token. `\xad` is the Unicode soft-hyphen (U+00AD). The soft hyphen is used to suggest where it makes sense to hyphenate a word if a line-break is needed. This shows up fairly frequently (2.9k occurrences) on GitHub in web-scraping datasets, which suggests that a model trained on data scraped from the web might see a fair number of these.

Basically, OpenAI is trained on web data that has a number of words where splitting on -der makes sense (e.g., mur-der, un-derstanding, won-derful; although the most common occurrence in a GitHub search for "\\xadder" is what appears to be an incorrectly encoded string "L\xc3\xadder", probably from the Portuguese and Spanish "Lí-der").

Anyways, using the o200k tokenizer `mur\xadder` yields two tokens (88762 and 179582). 88762 encodes "mur" and 179582 encodes "\xadder".

New comment by blochist in "How each pillar of the First Amendment is under attack"

blochist — Mon, 31 Mar 2025 17:25:44 +0000

And where in the Constitution do we find this notion of "rights of citizens". The Constitution governs how the federal government is allowed to act and the powers it has been given (by the states and people).

New comment by blochist in "HHS Grants Terminated [pdf]"

blochist — Wed, 26 Mar 2025 13:33:38 +0000

The NIH has been funding studies in China run by Chinese institutions since at least 1985 (see https://reporter.nih.gov/search/feqMQ1HUnUmWJHKzGl8wMA/proje...).