Hacker News: 8note

New comment by 8note in "How the Claude Code team does code review"

8note — Sun, 18 Oct 2026 19:38:32 +0000

i still havent seen claude write code as bad as legacy spaghetti, but i guess im never quite vibe coding since i usually have an idea of the architecture i want, and that leaks into my prompts about how it should approach the problem and verification of the solution?

New comment by 8note in "Zuckerberg's Meta Takes Down No Kings Instagram During Protest Day"

8note — Sun, 18 Oct 2026 03:04:51 +0000

i havent really seen proper breaches of free speech until those same people got power, like the government coming in to strip people's education visas or greencards for having the illegal opinions about israel

New comment by 8note in "Waymo in Toronto"

8note — Fri, 16 Oct 2026 21:36:39 +0000

i had a funny first experience where the lidar got super lost on a bush, but it did a uie around the block and pulled in from the other side, and from there the experimence was that it was the best driver ive ever seen

New comment by 8note in "US Rep Announces Congressional Legislation to Protect Video Game Ownership"

8note — Fri, 16 Oct 2026 20:52:55 +0000

no, poor people shouldnt have entertainment they've already paid for be taken away such that they have to spend a second or third time on it.

New comment by 8note in "Woman who rescued injured crow keeps getting 'thank-you gifts' from birds [video]"

8note — Fri, 16 Oct 2026 16:40:04 +0000

are you expecting her to eat the live crow?

the chicken is dead and cooked, but also has been raised with certain living standards and ensured to be a certain level of food safety

New comment by 8note in "New 'super ice' 10 times stronger, bringing it close to the strength of concrete"

8note — Fri, 16 Oct 2026 14:39:57 +0000

pykrete isnt a scifi material. its a ww2 idea the brits had for making an ice based aircraft carrier

New comment by 8note in "Containment and monitoring failures are to blame for AI hacking reports: experts"

8note — Wed, 14 Oct 2026 19:47:27 +0000

and yet regulating training data is not what's coming up with AI safety, even though it can solve that, and many other problems like copyrights

New comment by 8note in "Containment and monitoring failures are to blame for AI hacking reports: experts"

8note — Wed, 14 Oct 2026 19:39:28 +0000

1. hey chatgpt, find all the issues with this sandbox

2.hey chatgpt, fix this list of issues with the sandbox

3. goto 1

New comment by 8note in "Containment and monitoring failures are to blame for AI hacking reports: experts"

8note — Wed, 14 Oct 2026 19:38:23 +0000

some of the obvious steps to testing an agent i think would be to have it first test its own environment?

and then you fix the issues it found, and run your actual tests

New comment by 8note in "Containment and monitoring failures are to blame for AI hacking reports: experts"

8note — Wed, 14 Oct 2026 19:37:07 +0000

previously we described your definition for "rogue agent" as hallucinations - the operator expected a real answer to their question, and the agent went rogue and lied about the answer instead.

but also, ai has been going rogue for decades. the GPS machines of old went rogue all the time, routing drivers into lakes, rivers, and construction, sometimes with deadly results. The operator submitting an address wanted driving isntructions, but instead, the ai went rogue and sent them a bad route

New comment by 8note in "Containment and monitoring failures are to blame for AI hacking reports: experts"

8note — Wed, 14 Oct 2026 19:15:02 +0000

whats the training data that points to using hacking for solving math problems though? what makes those likely tokens?

New comment by 8note in "Satya Nadella says we should assume all AI models are 'compromised'"

8note — Sun, 11 Oct 2026 05:22:36 +0000

the spine here is going with a deeply unpopular guy's dumb opinions and getting the hate mail

New comment by 8note in "500B tokens later: Letting AI agents decompile a first-person shooter"

8note — Sun, 11 Oct 2026 04:48:08 +0000

no, thats still functional here. the bugs have to be the same, such that speed runs could still run correctly

New comment by 8note in "I would like the value of my home to rise, while my property taxes fall"

8note — Sat, 10 Oct 2026 20:36:34 +0000

you could rent out a room though?

but you also have a much improved quality of life as a result of that property value going up, because theres more desirable stuff around

New comment by 8note in "I would like the value of my home to rise, while my property taxes fall"

8note — Sat, 10 Oct 2026 20:31:39 +0000

theres also (3) those single family home neighborhoods require much more government maintained infrastructure for water and roads

New comment by 8note in "I gave Opus 5.5 one prompt and six hours to visualize Invisible Cities"

8note — Fri, 09 Oct 2026 22:23:53 +0000

it hasnt actually be driven down to 10min though.

the 30min-3h of actual coding time is brought down to 10min, with the reviews, testing, consensus building, tracking, decision making all still taking up most of a week.

New comment by 8note in "I gave Opus 5.5 one prompt and six hours to visualize Invisible Cities"

8note — Fri, 09 Oct 2026 22:17:24 +0000

this is a forum built around building your own companies, no?

New comment by 8note in "Iranian campaign planted fake articles in real U.S. publications using ChatGPT"

8note — Fri, 09 Oct 2026 17:45:27 +0000

propaganda is also value neutral on those definitions

New comment by 8note in "US imposes sanctions on ICC hours after former judge wins Nobel Peace Prize"

8note — Fri, 09 Oct 2026 15:17:28 +0000

that would happen regardless of if americas oligarchs support the GoP. its the american people that voted in trump and got him majorities in both the house and senate

New comment by 8note in "Claude Code’s suggested message feature: I think the real customer is the model"

8note — Wed, 07 Oct 2026 14:56:32 +0000

with more tokens it probably could be. you dont need the real tool call response, if the likely tokens after the tool call(s) is that the user is unhappy, it could predict all those tokens, then have a secomd model predict whether the tool call should be made or not