HackerTrans
トップ新着トレンドコメント過去質問紹介求人

hansonkd

1,125 カルマ登録 7 年前
CTO at getdynasty.com

コメント

hansonkd
·昨日·議論
Yeah, the flip side of the article is that Fable level models can fix the majority of codebases created from the past 3 years and one shot it to a fixable state that is "human maintainable"
hansonkd
·昨日·議論
The comments that drive the most homicidal behavior are outdated or inaccurate comments rather than no comments.
hansonkd
·3 日前·議論
to be fair, nobody said they were shocked. just that they were a bad trading partner.
hansonkd
·18 日前·議論
I guess Zelle is the replacement. I've sent thousands to strangers with it with zero problems or fees. I don't like it at all because it is private and in my view a way for banks to avoid accountability.

But also zelle isn't really an option if you are anything bigger than a single person business.
hansonkd
·25 日前·議論
2,000 gold bars were stopped out of iraq. and that is just what was stopped / reported.

Dictators and autocrats may or nay not have cash sitting in a bank account, but there are most likely multiple with $1B in gold.
hansonkd
·先月·議論
Not to be pedantic and I agree with what you are saying but:

> As a rule of thumb, if you look at something for 3 minutes and have some obvious questions, the scientists that looked at it for several years of their life in great detail might have had those same obvious questions as well

This does not mean that just because they had those obvious questions that they were properly resolved. Human history has a long track record of people who knew better but chose to ignore. In science there is an incredible pressure to have positive results rather than negative ones (IE nobody would care or know about this study if the title was "we looked and iphone doesn't explain 33-52% of fertility decline"
hansonkd
·先月·議論
GetDynasty.com | Engineering (In-Person San Fransisco or US Remote initially OK) | Junior Engineer | $100k-$120k + equity

Dynasty helps startup founders legally eliminate millions in capital gains taxes through QSBS trust planning. Today, sophisticated tax strategies are locked behind elite law firms and opaque processes. We’re building the software layer that makes advanced trust and tax infrastructure accessible, understandable, and fast.

We combine fintech, legal infrastructure, and thoughtful product design to turn what used to be months of paperwork and six-figure legal bills into a streamlined, guided platform. Our customers are venture-backed founders preparing for exits when the stakes are high and precision matters.

We’re early, fast-moving, and product-obsessed. This is a chance to help define the system of record for founder wealth.

Positions:

* Hiring a junior engineer with a passion for building to help create the next generation of financial platforms.

* Senior product focused builder with a passion for design and UX to join our team in San Fransisco.

* Senior engineer focused on agentic workflows and development.

email: [email protected]

Contact: [email protected]
hansonkd
·先月·議論
In general once or twice a month cleaners aren't hired to "tidy up", they deep clean.

a bit like the difference of brushing your teeth and going to a hygienist.
hansonkd
·3 か月前·議論
They almost certainly are, even if unknowingly, because HN and all blogs get piped continuously into all models' training corpus.
hansonkd
·3 か月前·議論
After the steep increase in sales of Mac Studios specifically for LLMs, I'm waiting for Apple to release a frontier level model, optimized for highest end of apple hardware (probably would be hardware locked by a certain neural processor needed (which would then lock the memory config).

The built in Apple Intelligence right now is very small, but even just having a small LLM you know is always there, online, fast and ready makes you think about building app differently. I would love the context to expand from the meager ~4K tokens.
hansonkd
·3 か月前·議論
bigger filter = less noise to move more air
hansonkd
·3 か月前·議論
Its actually simpler since the skill can be 100% a MD file.
hansonkd
·3 か月前·議論
idk, just have a standard internet request tool that skills can describe endpoints to. like you could mock `curl` even for the same CLI feel
hansonkd
·3 か月前·議論
I've been thinking a lot about this lately. It seems like what is missing with most coding agents is a central source of truth. Before the truth of what the company was building and alignment was distributed, people had context about what they did and what others did and are doing.

Now the coding agent starts fresh each time and its up to you to understand what you asked it and provide the feedback loop.

Instead of chat -> code, I think chat -> spec and then spec -> code is much more the future.

the spec -> code phase should be independent from any human. If the spec is unclear, ask the human to clarify the spec, then use the spec to generate the code.

What happens today is that something is unclear and there is a loop where the agent starts to uncover some broader understanding, but then it is lost the next chat. And then the Human also doesn't learn why their request was unclear. "Memories" and Agents files are all ducktape to this problem.
hansonkd
·3 か月前·議論
You know, that would actually be pretty fun and cool. Like if you had home automation set up with a "pet assistant", but it would only follow your commands if you made sure to keep it happy.
hansonkd
·4 か月前·議論
I mean emails were and still are a huge security risk. Sometimes I'm more scared of employees opening and engaging with emails than I am than anything else.
hansonkd
·4 か月前·議論
Apple is terrible for business. Every portal and product require a new apple id. apple store and apple business can't be same apple id. your device id can't be the same as either. Its madness. Last count i have 4 apple ids that I have to shuffle around.
hansonkd
·4 か月前·議論
Any of these abstractions are just temporary.
hansonkd
·4 か月前·議論
Floats break the basic expectation of == for round-trip verification, not due to programmer error, but because NaN is non-reflexive by spec. A bit-perfect round-trip can reproduce the exact bit pattern and still fail an equality check. The problem is intrinsic to the type, not the operator.
hansonkd
·4 か月前·議論
I love to build a plan, then cycle to another frontier model to iterate on it.