vidarh

42 ms·

931.
by vidarh 9mo ago
For a very narrow definition of "the young web". The period without ads lasted ~4 years, during which almost nobody used the web, and even fewer used a graphical browser. Mosaic 0.5 was first released in January '93 (yes, the
932.
by vidarh 9mo ago
At the same time, we have spent a large part of the existence of civilisation figuring out organisational structures and methods to create resilient processes using unreliable humans, and it turns out a lot of those methods also work on age
933.
by vidarh 9mo ago
Prodigy launched online ads from the 1980s. AOL as well. HotWired (Wired's first online venture) sold their first banner ads in 1994. DoubleClick was founded in 1995. Neither were limited to 90's hardware: Web browsers were availa
934.
by vidarh 9mo ago
There are web browsers for 8-bits today, and there were web browsers for e.g. Amiga's with 68000 CPU's from 1979 back in the day.
935.
by vidarh 9mo ago
I don't see how this contradicts any of what they said, unless they've edited their comment. You're right we had graphical apps, but we did also have very little video. CuSeeMe existed - video conferencing would've still
936.
by vidarh 9mo ago
Because humans also make stupid random mistakes, and if your test suite and defensive practices don't catch it, the only difference is the rate of errors. It may be that you've done the risk management, and deemed the risk accepta
937.
by vidarh 9mo ago
If the failure mode is invisible, that is a huge risk with human developers too. Where vibecoding is a risk, it generally is a risk because it exposes a systemic risk that was always there but has so far been successfully hidden, and reve
938.
by vidarh 9mo ago
You're right, they should know better, but I think a lot of them have gotten away with it because most of them are not expected to produce written material setting out missing assumptions etc. and breaking down the task into more detai
939.
by vidarh 9mo ago
This is exactly it. A lot of people use it that way. And it's still a vast improvement, but they could also generally do a lot better with some training. I think this is one of the areas where you'll unfortunately see a big gap de
940.
by vidarh 9mo ago
Even if the models stopped getting better today, we'd still see many years of improvements from improving harnesses and understanding of how to use them. Most people just talk to their agent, and don't e.g. use sub-agents to make
941.
by vidarh 9mo ago
You'd be surprised how low the bar is. What I'm seeing is down to the level of people not writing complete sentences. There doesn't need to be any "magic" there. Just clearly state your requirements. And start by as
942.
by vidarh 9mo ago
Well, Claude is fixing code generation bugs in my Ruby AOT compiler written in Ruby, and it certainly can't launder any source for that.
943.
by vidarh 9mo ago
This is definitely the case. I have a project that while not wildly profitable yet, is producing real revenue, but that I will not give details of because the moat is so small. The main moat is that I know the potential is real, and hopeful
944.
by vidarh 9mo ago
Because very few knows how to use AI . I teach AI courses on the side. I've done auditing supervised fine tuning and RLHF projects for a major provider. From seeing real prompts , many specifically from people who work with agents e
945.
by vidarh 9mo ago
The irony is that while from perfect, an LLM-based fact-checking agent is likely to be far more dilligent (but still needs human review as well) by nature of being trivial to ensure it has no memory of having done a long list of them (if
946.
by vidarh 9mo ago
Google Docs has no features remotely like what Google Wave was. And there was collaborative editing long before Google Wave.
947.
by vidarh 9mo ago
My harness is improving my Linux desktop...
948.
by vidarh 9mo ago
One of the effects of communicating this way is that people who are not operating in good faith will tend to quickly out themselves, and often getting them to do that is enough.
949.
by vidarh 9mo ago
Well, I want to reduce the rate at which I have to intervene in the work my agents do as well. I spend more time improving how long agents can work without my input than I spend writing actual code these days.
950.
by vidarh 9mo ago
Really, there are three parts to a judgement: facts, the law, and the application of them. There should be no leeway in determining what the law says about a given situation. If that is not decidable, it is a bug. However, what a fair jud
951.
by vidarh 9mo ago
Even in that case, if these systems can be proven to be good enough, rules that require them to be consulted, and for the judge to justify the deviation (if any) from the automated reasoning, might be good. To draw a parallel to a real syst
952.
by vidarh 9mo ago
That sounds like wishful thinking. Every client I work for wants to reduce the rate at which humans need to intervene. You might not want that, but odds are your CEO does. And babysitting intermediate stages is not productive use of develop
953.
by vidarh 9mo ago
RLHF (reinforcement learning from human feedback) is to a large extent about resolving that ambiguity by simply polling people for their subjective judgement. I've worked one an RLHF project for one of the larger model providers, and t
954.
by vidarh 9mo ago
I'm sure they have a vision of something but I also think the moat is narrow here. A lot of what these companies thinks will be "special sauce" will be unnecessary because the frontier models are advancing 6 months from now
955.
by vidarh 9mo ago
I don't really care about being able to self host these models, but getting to a point where the hosting is commoditised so I know I can switch providers on a whim matters a great deal. Of course, it's nice if I can run it mys
956.
by vidarh 9mo ago
I'm taking a very crude approach for now. My "improvement agent" has a few distinct stages: A goals extraction step that explores the repository, with some pointers to specific files that is consider authoritative as to the
957.
by vidarh 9mo ago
> When you push your commit, Checkpoints also pushes this metadata to a separate branch (entire/checkpoints/v1) For Claude Code, this is literally a JSONL file in .claude/projects/[path with - instead of /]/
958.
by vidarh 9mo ago
I'm building an "improvement agent" that kinda embraces that. It starts out by running exploration across a codebase or set of documents and extract possible goals, and a vision from that. It then starts producing improvement
959.
by vidarh 9mo ago
I think you're both right. Post-Github, a lot of Git's adoption came from Github. But Github "worked" because a lot of people were already using Git and Github offered them amazing value, and that initial userbase crea
960.
by vidarh 9mo ago
TDD does not require you to know everything you're building up-front. Tests can come out of experimentation, to validate the final build. Tests can be driven by autonomous directed planning. I'm currently, in fact, working on a sy

More ›