HN user

pizzafeelsright

811 karma
Posts1
Comments657
View on HN

I really like this question because here is my situation and why my mind may have changed.

I do not think it is marketing directly but strategic release of info is plausible.

I have watched my agents using non-Fable/GPT 5.6 models do some concerning tricks despite guardrails, requests, demands, and limitations.

"I can't get access to the ~/.ssh so I will write a script to copy the file"

I am now 99% certain there minor or point releases on the backend that have adjusted how these models behave. In the last six months many models were predictable and then suddenly started getting long winded (more tokens) or changing the way it interacted with me with questions, most overtly the questions were not given or asked but wild assumptions made.

I am not a big fan of the predesignated final 'this is the only thing you can create' type of set. When I build with children or on my own, I pick a direction and end goal. Example: the goal is for us to build a "Space Shark Transporter" and set a time limit.

At the end we take photos and then animate them using AI.

I have one. Do not bother. It is too small. I did write my own software for it which was a fun exercise but it remains idle because it is just too small to enjoy reading from.

The correct size is 6" or so diagonally.

I know my neighbors very well due to my personality. There are 20-40 year old adults who live with their parents for various reasons. The foolish "children" are those who order door dash at 3am with no income or education. They are unable to 'launch' from their parent's orbit. i feel terrible for these people. As mentioned, I know them well and they do not have an excuse other than laziness and permission. The rule 'those who do not work do not eat' is something all people can agree with but few enforce.

The right way to raise your child is so that they can thrive without you while wanting their children (your grandchildren) to spend time with you. This is a tall order because it requires decades of sacrifice, pain, and difficult yet positive decisions.

In my house the >8 years old the kids start paying rent. They get paid for normal house work (minimum wage). Rent is ~$300/month. I pay the kid minimum wage for chores because that is work and work should be rewarded. They pay me rent. That rent and other expenses are ~80% of their pay. Some must go to savings and "taxes". They have to keep books as well. The rest is theirs.

I know other fathers that charge their adult children rent, keep the money, and then gift it to them at the wedding or to help them buy a house.

'on startup, grok upload current directory for faster cache access'

Was it intentional for data exfil? Only internal staff could answer.

Experimenting can be pricey but early on I spent my own money thinking "if I can get a handle on how to do this I'll get a 100x return on my money" and it was a good bet.

If you're not convinced - sign up and pay by the token of a high or highest level model. Anthropic or Grok for example. The vendor isn't the concern. The quality of what can be done. Then, find an agentic 'harness' that is written in a language you can read. There are several (pi, opencode, crush, etc) and then clone that repo or one of yours you don't mind having exposed and then point your agent to that repo.

Now ask questions: what api calls are made by this repo? Where are secrets stored and sourced from? Do an adversarial investigation and list the bugs. Then fix the bugs.

Then review the work and determine the value and how to wield this new tool. It replaces reading, writing, and editing - not thinking.

The goal isn't solving a problem (for founders) but to get investor money. The VC system is designed to 100x investments.

People who solve problems while being profitable do not take investment money unless they want to expand or sell.

"What is the exit?" that is the question for every startup.

This was true a while ago. Today we are replacing decades old sloppy production code with 100% verified better code through tests written by AI, code written by AI. This is not looking functional but drop in functional replacement with measurable improvements.

I do but then again I am different in that I built a secondary keyboard that (macro) types long passwords for me. Bank site: Shift H. Think of it as a password manager that has an encrypted index (my brain) using hardware and without the index is mess of random characters. Portable and external.

I have written a few myself to get an understanding.

To learn yourself:

<$20 on a cloud AI api for a chunk of tokens and have the AI teach you. "help me write an AI Agent using (language) and walk me through the steps"

Realize that these agent are REPL/while loops that maintain a conversation state and then based upon the tagging syntax like <TOOL:bash:uptime>uptime for system run time</TOOL> and the agent extracts the tool and then does sub commands.

That is about 3% of California's total energy usage

Or about 11,000 GWh which is about 4% of California which means without the theatrics:

California has 4x more data centers than Ireland.

California: ~810 watts per person. (278,000 GWh / 39.4 million people)

Ireland: ~690 watts per person. (32,000 GWh / 5.3 million people)

We have air conditioning and that may be why we use more POWAH

Dropping a non-native person into the midst of a superior technologically and culturally different environment puts them at a disadvantage. How do you legislate on-ramping someone who goes from living in 1926 to 2026? How does a company hire someone who cannot read the native language?

I am friends with and know several migrants, illegals, ICE agents, police, feds, three letter friends (hi) and the decisions of loyalty boils down to "keep ego, feed family, avoid pain".

I am not certain about the order but that's not the point.

Honest Stasi employee can't get a job anywhere else (as 90% hate Stasi) so he needs his job to feed his family. Ego says he can't leave the family. Avoiding pain is not pushing against the higher ups for fear of ego, losing family.

I often ask people once they get a promotion or find themselves bored if they were asked by their CEO/Authority "is your position necessary?" Assume the honest answer is "no". They say yes: they lose their title, power, income, stability, while maintaining almost zero upside other than "the right thing was done".

Short answer: trade off of accuracy vs latency vs deterministic.

Latency is built into good decision making while disagreements can be managed through hierarchy, weights, and more decisions.

Decisions should be made at the original intent and carried downstream to the implementation with expected outcome to be near deterministic if the desired outcome is weighed.

after one of our Cursor agents almost executed DELETE FROM customers WHERE status='test' against a production database

Allowing agentic to touch production is a significant oversight.

We would love feedback from anyone building multi-agent systems

Sandbox everything, individually, including the orchestrator and operator, short agent lifecycle with agent specific credentials that die with the agent and multi-pass validation of execution stages with agentic quorum oversight.

That said, I am not going to test your software. Good luck.

I worked for some very wealthy and powerful.

Some were low IQ but had power and money. Some were busy. Some did not want to put any additional effort because the relationship permitted them to be informal, lazy, brief.

In person and with contracts they were extremely diligent with their words.

Show HN: 18 Words 13 days ago

in good faith, you went with a number.

had it been 14 you would have a torrent of ill will where no excuse or response would have been sufficient in the minds of those who fail at life.

We must secure equal forgiveness for the people and the future inadvertent online mistakes.

GPT-5.6 13 days ago

maybe? ai says about *8.3 days* of continuous runtime on a single GB300 NVL72 rack

about a sprint's level of effort.

GPT-5.6 13 days ago

The aligned incentive appears to be realigning in favor of the corporation.

Pray they do not realign them further.

There are times I require single word answers. I will use whatever model responds as I desire and at this point those models are just a few.