samuelknight

6 ms·

61.
by samuelknight 6mo ago
> On top of that, there are a lot of personal and subjective aspects to code. You might have certain preferences about formatting, style, structure, dependencies, and approach, and I have mine. Code formatting is easy to solve. You write
62.
by samuelknight 6mo ago
I like the website. Simple and snappy.
63.
by samuelknight 6mo ago
The cost of intelligence is non-linear, with slightly dumber models costing much less. For a growing surface of problems you do not need frontier intelligence. You should use frontier intelligence for situations where you would otherwise re
64.
by samuelknight 6mo ago
I don't know about Mythos but the chart understates the capability of the current frontier models. GPT and Claude models available today are capable of Web app exploits, C2, and persistence in well under 10M tokens if you build a good
65.
by samuelknight 6mo ago
All code is a liability in general. All code written before LLMs and during the current in-between years are vulnerable to the next frontier model. Eventually we will settle into a new paradigm that correctly addresses the new balance of ef
66.
by samuelknight 6mo ago
Have you considered poking the cache? When a user walks away during the business day but CC is sitting open, you can refresh that cache up to 10x before it costs the same as a full miss. Realistically it would be <8x in a working day.
67.
by samuelknight 7mo ago
On the one hand I don't understand why it needs to be half a million lines. However code is becoming machine shaped so the maintenance bloat of titanic amounts of code and state are actually shrinking.
68.
by samuelknight 7mo ago
This is a physical implementation of a tiered caching hierarchy.
69.
by samuelknight 7mo ago
This is standard in every tech RSU vest schedule I have seen.
70.
by samuelknight 7mo ago
I think that a model designed to ignore semantic chatter like financial news and deeply inspect the raw data is a very powerful perspective.
71.
by samuelknight 7mo ago
Large companies already maintain a clone of their packages. Very large ones actually bundle their own build system (Google Bazil, AWS Brazil). If you want to update a package, you have to fetch the sources and update the internal repository
72.
by samuelknight 7mo ago
Yes, that's how they become customers.
73.
by samuelknight 7mo ago
Absolute wave of supply chain attacks recently. Hopefully this causes everyone to tighten up their dependencies and update policies.
74.
by samuelknight 7mo ago
Ridiculous string comparisons on long chains of logic are a hallmark of vibe-coding.
75.
by samuelknight 7mo ago
If that was a jab it my writing then yes, I am absolutely being sincere because I am an expert on this topic. LLMs went from being ok at one-shoting a function a to being so good at hacking that it's difficult to evaluate them. Prospec
76.
by samuelknight 7mo ago
LLMs are expert hackers because: 1) They are expert coders, including a decently comprehensive CVE knowledge 2) They know every programming language/framework/stack 3) They know every human language They already have super human b
77.
The new security frontier for LLMs; SIEM evasion(blog.vulnetic.ai)
2 points by samuelknight 7mo ago|1 comments
78.
by samuelknight 7mo ago
What does Valve ship without x86?
79.
by samuelknight 7mo ago
One of Terraform's most powerful features that it will tell exactly which resources change before it makes the changes. The hard part is writing Terraform, not reviewing and running one command. In my workflows I am the one who runs &q
80.
by samuelknight 8mo ago
Copycats are not a new problem. You can be completely open source and have a trademark on the project name.
81.
by samuelknight 8mo ago
LLMs and particularly Claude are very capable security engineers. My startup builds offensive pentesting agents (so more like red teaming), and if you give it a few hours to churn on an endpoint it will find all sorts of wacky things a huma
82.
by samuelknight 8mo ago
https://metr.org/blog/2025-07-10-early-2025-ai-experienced-o... That info is from mid 2025, talking about models released in Oct 2024 and Feb 2025. It predates tools like Claude Code and Codex, Lovable was 1/3 cur
83.
by samuelknight 8mo ago
There simply isn't enough popcorn for the fast AGI timeline
84.
by samuelknight 8mo ago
I approve of this interaction because squabbling with bots is funny. Make no mistake that in human society it's the humans that run the show even when the bots bring out their most persuasive arguments like 'bigotry' and &#x2
85.
by samuelknight 8mo ago
This is what I expect from my employees
86.
by samuelknight 8mo ago
Someone still has orchestrate the shit show. Like a captain at the helm in the middle of a storm. Or you can be full accelerationist and give an agent the role of standing up all the agents. But then you need someone with the job of being a
87.
by samuelknight 8mo ago
I the article is wrong. UNLOGGED means it isn't written to WAL which means recovery and rollback guarantees won't work since the transaction can finish before the page can be synchronized on disk. The table loses integrity as a tr
88.
by samuelknight 8mo ago
Skeptical about replacing Redis with a table serialized to disk. The point of Redis is that it is in memory and you can smash it with hot path queries while taking a lot of load off the backing DB. Also that design requires a cron which mea
89.
by samuelknight 9mo ago
Open models have been about 6 to 9 months behind frontier models, and this has been the case since 2024. That is a very long time for this technology at it's current rate of development. If fast takeoff theory is right, this should wid
90.
by samuelknight 9mo ago
You understate the capabilities of the latest gen LLMs. I can typically describe a user's bug in a few sentences or tell Claude to check fetch the 500 error in Cloud run logs and it will explain the root cause, propose a fix, and throw

More ›