You should really only get one pass for human rights violations before skepticism is the default response. Facebook already used theirs on genocide. After something like that, why should the public give them any benefit of the doubt?
(DBOS co-founder here) From a DBOS perspective, the biggest differences are that DBOS runs in-process instead of on an external server, and DBOS lets you write worklflows as code instead of explicit DAGs. I'm less familiar with Hatchet, but here's a blog post comparing DBOS with Temporal, which also uses external orchestration for durable execution: https://www.dbos.dev/blog/durable-execution-coding-compariso...
Yep, durable execution-wise we're targeting a very similar use-case with a very different philosophy on whether the orchestrator (the part of the durable execution engine which invokes tasks) should run in-process or as a separate service.
There's a lot to go into here, but generally speaking, running an orchestrator as a separate service is easier from a Postgres scaling perspective: it's easier to buffer writes to the database, manage connection overhead, export aggregate metrics, and horizontally scale the different components of the orchestrator. Our original v0 engine was architected in a very similar way to an in-process task queue, where each worker polls a tasks table in Postgres. This broke down for us as we increasing volume.
Outside of durable execution, we're more of a general-purpose orchestration platform -- lots of our features target use-cases where you either want to run a single task or define your tasks as a DAG (directed acyclic graph) instead of using durable execution. Durable execution has a lot of footguns if used incorrectly, and DAGs are executed in a durable way by default, so for many use-cases it's a better option.
Hatchet looks very cool! As an interested dilettante in this space, I’d love to read a comparison with Dagster.
Re DBOS: I understood that part of the value proposition there is bundling transactions into logical units that can all be undone if a critical step in the workflow fails - the example given in their docs being a failed payment flow. Does Hatchet have a solution for those scenarios?
The core idea being that you write the "parent" task as a durable task, and you invoke subtasks which represent logical units of work. If any given subtask fails, you can wrap it in a `try...catch` and gracefully recover.
I'm not as familiar with DBOS, but in Hatchet a durable parent task and child task maps directly to Temporal workflows and activities. Admittedly this pattern should be documented in the "Durable execution" section of our docs as well.
Re Dagster - Dagster is much more oriented towards data engineering, while Hatchet is oriented more towards application engineers. As a result tools like Dagster/Airflow/Prefect are more focused on data integrations, whereas we focus more on throughput/latency and primitives that work well with your application. Perhaps there's more overlap now that AI applications are more ubiquitous? (with more data pipelines making their way into the application layer)
My wife asked the same question when I was restoring the GBA I found in the closet, so I set up an emulator with the ROM on her phone. Less storage space than most apps and doesn't need ads or tracking.
Lost Frontier on mobile is good. Better than Wargroove imo. It really feels like the original Advance Wars in its simplicity. Just in a Western cowboy setting.
I’ve had Warside wishlisted on Steam for a year now, as it looks to be just that. And when I looked it up just now to double-check the name, looks like it’s coming out next month! Fingers crossed.
(DBOS co-founder here) DBOS embeds durable execution into your app as a library backed by Postgres, whereas Restate provides durable execution as a service.
In my opinion, this makes DBOS more lightweight and easier to integrate into an existing application. To use DBOS, you just install the library and annotate workflows and steps in your program. DBOS will checkpoint your workflows and steps in Postgres to make them durable and recover them from failures, but will otherwise leave your application alone. By contrast, to use Restate, you need to split out your durable code into a separate worker service and use the Restate server to dispatch events to it. You're essentially outsourcing control flow and event processing to the Restate server., which will require some rearchitecting.
Any reason why you're calling out a need for a framework to be js/ts based? There's plenty of python frameworks in active development, some of which have js bindings/libraries.
minimal setup. Anything python requires the host machine to have the correct python version, pm and libs installed (which is far more than normal users can do), or have it compiled within a virtual python executable (big!).
Web-native technology requires minimal setup as it can basically run in any browser (or electron) as is.
uv has revolutionized the Python situation, mostly.
I recently updated YOShInOn's Python environment to be repeatable, the python packaging part was pretty simple, but getting CUDA running in WSL2 was a little tricky. Turns out my "game ready" NVIDIA drivers in Windows install a certain version of the base CUDA libs in WSL somehow. You have to install another 5 libraries with deb packaging inside the WSL which is not too hard but I realized I had a version mismatch about 1/3 of the way through but decided to barrel ahead. I installed most of the deb's manually but got fatigued and installed the last one automatically. Somehow it all works so I'm not messing with it, but I am a bit intimidated about what to do if I have problems and need to tear it down.
uv is still a cli tool, maybe ok for tech enthusiasts. No normal user will ever install a cli tool. And even for knowledgable users like yourself, there are a manifold of different problems as you stated.
It can be called from another application and hidden, probably a mostly Java application could uv up a Python environment for one or more subsystems.
My beef is with CUDA being not just one thing but about 5 libs if you wanna use PyTorch. In the old days I figured how to make conda packages for all that NVIDIA stuff so I could single step install any version of Tensorflow and it was 100% correct. Just then they changed Tensorflow so it was 95% correct but worked most of the time for most people and my system didn’t work anymore.
What they did is illegal. Any rank and file that did the same would be in prison for a decade, no questions asked.
In general, it seems like you're trying to "3d chess" incompetence into strategy, but try taking a step back and looking at it with clear eyes. This was a bad decision, plain and simple. Nobody is taking responsibility for it, and that makes it worse - these people are in charge of the largest intelligence and war machine on the planet. This is not okay.
The reality, which people are not acknowledging here, is that what they did may not have been according to official policy but it has been normal and pervasive for decades. It isn’t partisan, everyone does it. This is how DC works and the American public just got an education.
As a consequence, any enforcement now would be viewed as extremely selective.
I have been exposed to a lot of classified information in meetings in DC that were supposed to be unclassified. This isn’t an isolated incident, it has been a systemic issue across every administration for as long as I’ve worked in DC.
People should focus less on the incident and more on why this has been normal for decades.
The underlying tension is that doing things the official way is extremely slow and speed matters. There is a longstanding bias toward taking more risks in terms of information exposure because being slow carries its own significant risks. Speed of decision making is critical and that has proven to be impossible if every interaction has to happen inside a SCIF. It is a tension the intelligence community is still grappling with.
You're embarrassing yourself, brother. Nobody is asserting that this is OK. It was naive to assume the government was secure in the first place. Privacy advocates and whistleblowers have whistleblowers have been saying this for decades! You just weren't paying attention. WikiLeaks and Snowden leaks wasn't a "fun" news cycle, it was revealing everything you need to know about how the government operates truly. With no concern for security
Of course they haven't. Every think-tank moron knows political opsec is a joke (this is why sigint works in the first place) let alone people actually working in politics
I'm not doing anything of the sort. The kind of problem I'm flagging in is experienced every day by governments all over the world. Would anyone disagree? People on here who want to put their heads in the sand about it are just being political when there is a legitimate technical topic to discuss. The point is these aren't "rank and file" actors. They are at the top of political leadership. Those rules don't apply at this level of power politics so why get bogged down in such thinking?
Because laws should matter. Laws should apply to members of government too. Unless you're suggesting it's totally fine for Trump and his administration to be above the law. In which case the whole discussion is moot, because then it's not a democracy with a functioning rule of law anymore.
Law is a tool, and some tools are appropriate for some contexts and others are not. Do you think there is such a thing as "International Law"? If so, I would ask you what you think that actually is and where its legitimacy comes from and who enforces it? Politics and Law are two separate spheres of human conflict. You actually degrade the law by trying to weaponize it for political purposes. I would hope the past 10 years have shown that to everyone.
> What they did is illegal. Any rank and file that did the same would be in prison for a decade, no questions asked.
IIUC, the "rank and file" go to prison for violating their NDA. At the highest level these people are appointed and don't have an NDA which is why senators / representatives can leak without punishment.
This is something I wanted but I couldn't figure out a way to do it in a way that's meaningful. Authors like Simon Willison publish frequently, so even though he has a lot of high-scoring posts, he has a lot of low-to-no-scoring posts too. It feels unfair to penalize people who publish frequently just because not every post is a homerun.
Those 3x price markups on fast food aren't going to pay for themselves, you know.
reply