Can someone explain what value openclaw provides over like claude code? It seems like it's literally just a repackaged claude code (i.e. a for loop around claude) with a model selector (and I guess a few builtin 'tools' for web browsing?)
From what I remember, the key differentiating features were:
- a heartbeat, so it was able to 'think'/work throughout the day, even if you weren't interacting with it
- a clever and simple way to retain 'memory' across sessions (though maybe claude code has this now)
- a 'soul' text file, which isn't necessarily innovative in itself, but the ability for the agent to edit its own configuration on the fly is pretty neat
Its a coding agent in a loop (infinite loops are rejected by coding agents usually) with access to your computer, some memory, and can communicate through telegram. That’s it. It’s brilliant though and he was the first to put it out there.
I see, so there's actually an additional for loop here, which is `sleep(n); check_all_conversations()`, that is not something claude code does for sure.
As far as the 'soul' file, claude does have claude.md and skills.md files that it can edit with config changes.
One thing I'm curious about is whether there was significant innovation around tools for interacting with websites/apps. From their wiki, they call out like 10 apps (whatsapp, teams, etc...) that openclaw can integrate with, so IDK if it just made interacting with those apps easier? Having agents use websites is notoriously a shitty experience right now.
The main one is that you can run and/or host it remotely, unlike Claude Desktop. By this I mean, you can run OpenClaw on a service like Tailscale and protect your actual machine from certain security/privacy concerns and - regardless of the choice - you can connect your access to OpenClaw via any chat agent or SSH tunnel, so you can access it from a phone. If Claude Cowork comes to iOS/Android with a tunnel option, they can resolve this difference.
A smaller difference would be that you can use any/all models with OpenClaw.
Hmm, whats stopping you from running claude code on a separate machine you can ssh into? I don't understand that point at all, I do that all the time.
Using a claude code instance through a phone app is certainly not something that is easy to do, so if there's like a phone app that makes that easy, I can see that being a big differentiator.
Something I learned while hacking on something recently is that claude’s non-interactive mode is super powerful. It uses all the same tools/permissions etc as interactive would, it can stream responses as JSON with tool use etc, and it can resume from a previous session. Together this means you can actually build a really slick chat-like UI for it.
This is using Claude on VMs that don’t have SSH, so can’t use a regular terminal emulator. They stream responses to commands over websockets, which works perfectly with Claude’s streaming. They can run an interactive console session, but instead I built a chat UI around this non-interactive mode.
OpenClaw is probably overkill if you just want to have a nice remote UI to access claude code, do tool call approvals. There are a ton of remote cli apps and guides to setup ssh access via tailscale etc, but none that just work with a nice remote web interface.
For me personally I can't stand interacting with agents via CLI and fixed width fonts so I built a e2e encrypted remote interface that has a lot of the nice UI feature you would expect from a first class UI like Claude Vscode extension (syntax highlighting, streaming, etc). You can self host it. But it's a little no dependencies node server that you can just npm install (npm i -g yepanywhere)
For programmers or people who know computers quite well the difference to claude code is small i would say. But for "Normies" its magical that you can just ask your computer to do anything from anywhere (set timers, install stable diffusion, send you a specific doc in your download folder). You don't even have to write it, you can send it a voice message and it will install whisper or send it to the openai whisper api, etc. Obviously this is more then dangerous, but looking at what passwords people still choose today (probably also the reason why everything requires MFA nowadays), most people don't care about Security.
Just uses claude. I haven't tried it much but it seems to be what you're describing.
Openclaw uses pi agent under the hood. Arguably most of the codebase could be replaced by systemd if you're running on a VPS for scheduling though, and then its a series of prompts on top of pi agent.
They serve different purposes. OpenClaw is supposed to be more of an autonomous sidekick assistant thing that can take instructions over different messenger channels. It can also be set up to take ongoing instructions and just churn on general directions.
> Crapping something out as quickly as possible and leaving somebody else to deal with the fallout of a bad data model and violent on-call isn't something to be rewarded IMO.
Sadly you've described precisely the optimal engineering strategy for promotion at my FAANG
And yet those five companies are among the most valuable in the world.
There's a cognitive dissonance that arises when you join a company that is performing extraordinarily well only to perceive dysfunction and incompetence everywhere you look.
It's so hard to reconcile the reality that companies can be embarrassingly wasteful, political, and arbitrary in how they run and yet can still dominate markets and print money hand-over-fist.
By being in the right place and right time once, making it impossible for users to leave, and buying up all their competition before they become a serious threat.
Also, they typically started with much more of a high quality culture when they were small and people's contributions were legible.
Once it turns into a giant bureaucracy with people you've never met judging a promo packet by rubrics, while they're unfamiliar with your whole org.. the incentives get diffierent.
People succeed in spite of these systems. They have resources, tremendous network advantages, and the people at the very top crust of engineers are indeed quite good at their job.
nothing says its the people getting promotions that are making the value.
there's still plenty of people not on that grind trying to make whatever thing nearby them work, or have other career goals than promotion or money
that promo option getting people who want to build a bunch of junk out of the way and into positions where they arent building stuff might be relevant to why those companies are succeeding
Because big companies can crush competition, either via lobbying for government regulations, acquiring the competitors, or driving the competition out of business by offering something comparable but cheaper or free.
It's the old Microsoft playbook of Embrace, Extend, Extinguish, but with more finesse.
It is also why their acquisitions tend to just die, because once the big company inefficiencies get integrated, the acquired startups just cannot function.
This is my every day, and I’m not at a FAANG. It is horrifying and infuriating at how much could be improved, and how much cost could be cut like that if there wasn’t an absurd amount of internal politics, and we had a high-trust environment.
Honestly, I feel like I have to know more and more these days, as the ais have unlocked significantly more domains that I can impact. Everyone is contributing to every part of the stack in the tech world all of a sudden, and "I am not an expert on that piece of the system" no longer is a reasonable position.
This is in tech now, were the first adopters, but soon it will come to other fields.
To your broader question
> Something that I think many students, indeed many people, struggle with is the question "why should I know anything?"
You should know things because these AIs are wrong all the time, because if you want any control in your life you need to be able to make an educated guess at what is true and what isn't.
As to how to teach students. I think we're in an age of experimentation here. I like the idea of letting students use all tools available for the job. But I also agree that if you do give exams and hw, you better make them hand written/oral only.
Overall, I think education needs to focus more on building portfolios for students, and focus less giving them grades.
> and "I am not an expert on that piece of the system" no longer is a reasonable position
Gosh that sounds horrifying. I am not an expert on that piece of system, no I do not want to take responsibility for whatever the LLMs have produced for that piece of system, I am not an expert and cannot verify it.
These are impressive metrics, are you able to make a living off of your 10M views?
I'm planning to leave my job this year and focus on content, mostly have been considering YouTube, but if blogging can work too, might consider that as well
Not even close to making a living! It does pay for my server though which costs $15 a month. YouTube gives you much more visibility. I'll try to compile the numbers from my single Carbon ad placement and the donations I receive from readers.
But I also don't think I have the process in place to do Blog, YouTube, Podcast and hold a full time job. Yes the job is my source of income.
Yeah I hear you. My understanding is that on youtube you can make ~2k per 1M views with the default ads. I'm hoping that I can be funded by some combination of that and something like patreon/membership/merch. But we will see, it's something I've wanted to do for years and I am getting too old to put off longer.
If you work at a megacorp right now, you know whats happening isn't people deciding to use less libraries. It's developers being measured by their lines of code, and the more AI you use the more lines of code and 'features' you can ship.
However, the quality of this code is fucking terrible, no one is reading what they push deeply, and these models don't have enough 'sense' to make really robust and effective test suites. Even if they did, a comprehensive test suite is not the solution to poorly designed code, it's a band aid -- and an expensive one at scale.
Most likely we will see some disasters happening in the next few years due to this mode of software development, and only then will people understand to use these agents as tools and not replacements.
...Or maybe we'll get AGI and it will fix/maintain the trash going out there today.
Also worked with Soumith. The man is a legend, moves mountains and completely changed the course of my career because he liked something I wrote. No arrogance, no politics, just an extremely down to earth and chill guy who elevates everyone around him.
Just went through that thread, I can't believe this wasn't a team of like 20 people.
It's crazy to me that apple would put one guy on a project this important. At my company (another faang), I would have the ceo asking me for updates and roadmaps and everything. I know that stuff slows me down, but even without that, I don't think I could ever do something like this... I feel like I do when I watch guitar youtubers, just terrible
I hope you were at least compensated like a team of 20 engineers :P
Sometimes (often?), one very dedicated and focused person is better than a team of 20+. In fact companies would do well to recognize these situations and accommodate them better.
> Back then, Apple had a sabbatical program that encouraged (mandated?) employees to take six consecutive weeks off every five years.
This is really a good take. I can't imagine companies give sabbatical programs nowadays. You still have your vacations so JK took 12 weeks (OP mentioned in the same comment). It was a boon for any system programmer who needs to clear his mind or deepen his thoughts.
This is amazing. I wonder what it took to port MacOS from PowerPC to Intel. Every assembly language part must be rewritten, that’s for sure. Anything else?
Didn't Nextstep support x86 long before MacOS X was a thing? I assumed that they always had it compilable on x86 long before the switch (since Rhapsody supported it). I guess the user space stuff might have been trickier but probably not the kernel itself and surrounding components.
Yeah but from what I read from the Quora answer, it sounds like JK did it from scratch? I could be wrong though. I just wonder how much effort is supposed to be put into such a project.
Likely a few foundational technologies that have had significant improvements/reimplementations from Rhapsody like the scheduler/threading infrastructure, memory management, Quartz, Carbon, Quartz.
How can you create a 'pressure differential' without deflecting some of the air away? At the end of the day, if the aircraft is moving up, it needs to be throwing something down to counteract gravity. If there is some pressure differential that you can observe, that's nice, but you can't get away from momentum conservation.
The pressure differential is created by the leading edge creating a narrow flow region, which opens to a wider flow region at the trailing edge. This pulls the air at the leading edge across the top of the wing, making it much faster than the air below the wing. This, in turn, creates a low pressure zone.
Air molecules travel in all directions, not just down, so with a pressure differential that means the air molecules below the wing are applying a significant force upward, no longer balanced by the equal pressure usually on the top of the wing. Thus, lift through boyancy. Your question is now about the same as "why does wood float in water"?
The "throwing something down" here comes from the air molecules below the wing hitting the wing upward, then bouncing down.
All the energy to do this comes from the plane's forward momentum, consumed by drag and transformed by the complex fluid dynamics of the air.
Any non-zero angle of attack also pushes air down, of course. And the shape of the wing with the "stickiness" of the air means some more air can be thrown down by the shape of the wing's top edge.
You can't, but you also can't get away from a pressure differential. Those things are linked! That's my main point, arguing over which of these explanations is more correct is arguing over what exactly the shape of an object's silhouette is: it depends on what direction you're looking at it from.
reply