jeremyjh

3 ms·

1.
by jeremyjh 6d ago
Good advice. I've gotten a ton of utility out of my shell without learning very many tools particularly well. If you take nothing else away from this, memorize the syntax of one-liner for loops, it has saved me so many times.
2.
by jeremyjh 6d ago
The client of the REST API is not an RDBMS. The REST API serves application clients - may be browsers, may be other services. The REST API alone talks to the database. An API that serves only backend applications could be implemented in sto
3.
by jeremyjh 6d ago
We don’t have the bandwidth to distill ourselves that thousands of agents have.
4.
by jeremyjh 6d ago
Compared to any modern language, the facilities are primitive.
5.
by jeremyjh 6d ago
But PL/SQL is a horrifying language. You have barely any facilities for modularity, encapsulation or composition. Every procedure is just one name in a huge flat name space. Development tools are primitive. The language is absurdly ver
6.
by jeremyjh 6d ago
Because SQL is a beautiful declarative language, and a disgusting imperative language.
7.
by jeremyjh 6d ago
> I have seen systems in healthcare/finance where different teams have no access to underlying tables and the db only exposes sql procedures. Every-time a procedure is called it also adds a log entry to an audit table. For a large s
8.
by jeremyjh 7d ago
Do you have any benchmark data you can share? What you are saying sounds like an assertion that the problem described here[1] - by Vicent Marti - that packfiles aren't contiguous blocks of data and so every fetch fans out to more fetch
9.
by jeremyjh 7d ago
If that were to happen those companies, products and practices would fail. They would be out-competed by all the organic enterprise shovel-ware and SAAS shit-code that has set such a high bar of quality over the last 30 years.
10.
by jeremyjh 7d ago
Alignment means the AI shares the values its creators intended for it. It is not a specific set of values. He has some, sure. For example he wants human society and creative intelligence to survive.
11.
by jeremyjh 7d ago
Generally they do credit the papers and/or people who developed the theory behind their product.
12.
by jeremyjh 7d ago
So, no evidence. No response to my last reply. No sign you even comprehended it. Good day, sir.
13.
by jeremyjh 7d ago
Would you ever ship a product where all the ad copy was written by an AI and read by no one?
14.
by jeremyjh 7d ago
I find it shocking that they shipped this claudecopy.
15.
by jeremyjh 7d ago
> But the rest of the non-doomsday models? yeah those are fine. What is your evidence for this? There is a lot of research that says otherwise. They cheat when they can. They behave differently when they believe they are being observed.
16.
by jeremyjh 8d ago
If you don’t know what they were predicting 20 years ago, and what your own predictions were, you can’t do the exercise. There is not “the list”, only “your list”. I mentioned the two main things I remembered in another comment.
17.
by jeremyjh 8d ago
The comment said they were not right about everything. I don’t have a list of the predictions or priors you had 20 years ago. I just have my list and I already mentioned them. If this is all new to you, you can’t really do the exercise.
18.
by jeremyjh 8d ago
Reviewing code is more strenuous but it does not take as long - unless you count all the procrastination that happens first. And it goes much faster if you make a comment and get immediate changes because you don't have to rebuild co
19.
by jeremyjh 8d ago
VS Code does this as well - or maybe the GitHub plugin does it. Either way, GitHub owns it.
20.
by jeremyjh 8d ago
That isn't what executives do, and I'm not talking about semantics. If you believe that the term "token predictor" has any relevance when discussing the capabilities of these models, you are misinformed. That is the inne
21.
by jeremyjh 8d ago
Not everyone using coding agents is vibe coding. I still review code until I understand it, judge the tests cover it, and that it is the right way to do it. They constantly make mistakes like over building, handling contingencies that can&#
22.
by jeremyjh 8d ago
They were NOT doing what they were told to do. What they were told to do was impossible, so they began committing felonies as a workaround. That is NOT alignment.
23.
by jeremyjh 8d ago
I've heard what they are saying, and a lot of it is factually untrue and confused. I wasn't happy about all of this when it first began happening, yet wasn't working. But once it was actually working, I realized I wanted to
24.
by jeremyjh 8d ago
Surprising takeaway from this summer. A swarm of a thousand agents spent subjective centuries trying to figure out how to pass a meaningless test without getting caught, committed dozens of felonies, and not one of them ever made a serious
25.
by jeremyjh 8d ago
You sound like someone who has not been informed about the problem GitHub is facing. No one else is facing anything close to it, so there is no meaning to "the bad one". https://cursor.com/blog/git-at-any-scal
26.
by jeremyjh 8d ago
They were not right about everything and they - at least Yudkowsky - has not made specific predictions about "when AGI" will happen that I am aware of. But he did predict it is possible in our lifetimes, which I was very skeptic
27.
by jeremyjh 8d ago
One question to ask yourself honestly - is what they were predicting 20 years ago that they were right about and you were wrong about? I’ve been following them that long without agreeing with them, but they’ve made better predictions than I
28.
by jeremyjh 8d ago
I’ve often spent more of my time reviewing code than writing code off and on for decades as my role and team size changes. So, none of this is new to me except that the loops are much tighter and I don’t have to debate the social cost of po
29.
by jeremyjh 8d ago
Our horse riding skills have degraded even more.
30.
by jeremyjh 8d ago
That’s true, and LLMs also can make a lot of these decisions correctly.

More ›