omniledger.io
I have built the first async, consistent data-platform
https://omniledger.io/
I also write about distributed systems
https://medium.com/@andrasgerlits
279 posts
128 followers
200 following
Regular Contributor
Active Commenter
comment in response to
post
To clarify: I would lose my _UK based_ contract and prospects, as all the banks I ever worked with expect a Certificate of Good Conduct issued by the Hungarian Government. If they stopped accepting what the Hungarian government has to say about my character, that would help.
comment in response to
post
See for yourself, I wouldn't believe it either
x.com/lone_rides/s...
comment in response to
post
Are the Count and the Swedish Chef DEI hires, do you think? I mean, I'm not sure how they got their jobs, when neither speak very good English. I'm sure there would be plenty of qualified muppets out there for the job, but nooo, they had to pick these two.
comment in response to
post
Arguably, that the SQL-spec just lumps all "write-anomalies" under "write-skew" is the culprit here, but facts remain, this would be a compliant implementation of the specification.
comment in response to
post
It would be interesting to think what would happen if such a lock-free transaction would "cross paths" with a write-lock, but I think the answer is that they would have to wait on its monitor and then be applied at a (basically random) order.
comment in response to
post
From where I'm sitting, it looks like Microsoft came up with this idea of "lost update anomaly", which would explain the need for locks, but that's not what the spec says.
comment in response to
post
I can't train the AI to do anything because it doesn't understand cause and effect. Like the cause being American and the effect being a sense of superiority even when your big new techs were invented by bottle-cappers?
Come to think of it, I now understand why Yanks are so hot on LLMs
comment in response to
post
Counterpoint: AI is a religion at this point, littering is bad.
comment in response to
post
This feels very much like the end-game. The last weeks of crypto looked very similar. Insane, pointless ads, purely speculative valuations with no products with wide adoption materialising even after years and absolute detachment from its tech-origins, all driven by cheerleaders.
comment in response to
post
Biden clearly didn't let Israel do whatever they wanted, which was why Netanyahu preferred Trump. Biden might not have done enough in your eyes (I accept this comes from a moral position, I'm not arguing ethics), but Trump was clear about the carte blanche. It's just a fact.
comment in response to
post
These look exactly like late-stage crypto commercials.
comment in response to
post
I don't know, saying there's no objective reality independent of observers to me means physics can't describe objective realities. Maybe this is a trivial result for everyone else, but it's certainly new to me
comment in response to
post
Useless? No. Overhyped? Certainly.
comment in response to
post
Well, the only solution that looks coherent to me (and apparently a number of others also) is that there's no "observer-independent" space-time, not sure what this actually means.
comment in response to
post
I've worked out the way to have linearly scalable, consistent, latency-mitigating data, potentially a single global SQL database. I would have never gone down this route without a patent. I sunk 10 years into this by now.
comment in response to
post
I've seen patents reaping serious benefits, if not building monopolies.
comment in response to
post
This will clearly be illegal in most of the EU, so I'm guessing this is a US-only thing?
comment in response to
post
This kind of talk was dominating most of our projects in the late 90s, when I started out. They weren't all wrong. Our projects often did reduce headcounts substantially in the end.
comment in response to
post
No, but I'm not on the committee drafting these.
comment in response to
post
This doesn't change facts. Homosexuality used to be classified as mental illness. If there's a list, a decision needs to be made if something makes the list or not, which requires consensus on the codifier's part. Therefore, in the end, this is a political question.
comment in response to
post
I'll stay factual. Mental illness is defined by the DSM and pretty much the whole Western world follows its definitions. This is how "gender identity crisis" was reclassified from mental illness to "gender dysphoria". I personally agree with this decision, but I understand many don't.
comment in response to
post
I understand that this is not helpful in this specific context, and that this not what people mean when they say "event driven", but strictly speaking, all locking is based on optimistic compare and swap, so you can always do locking asynchronously. I have a whole platform that does this.
comment in response to
post
Not even that. We know it doesn't work for what they want to use it for.
comment in response to
post
The business model of Airbnb aims to capture as much of the value pre-IPO as possible, not deliver a long-term sustainable business. People believe in the latter less and less.
comment in response to
post
It's such a sad state of affairs when stitch and fivetran qualify as central technologies. The entire industry just gave up on being innovative because they were told that's the best that can be done. It's just the death of ambition
comment in response to
post
I've seen plenty of fiction that were past their expiry date.
comment in response to
post
Knowing AWS will be your generation's ARPANET
comment in response to
post
No. Code is data that is being made sense of by the computer, which would be stateful in a naive interpretation. The code is totally passive. Whether the computer is an observer or not is I think a better question, but I also think it leads to QM like confusion around consciousness
comment in response to
post
A piece of stateless behaviour is data, being made sense of by the container its running in, which means the container is stateful, so is an observer of the data of the code. If the container couldn't redeploy or destroy, it would be stateless hardware.
comment in response to
post
I don't think a mirror is an observer.
comment in response to
post
Basically, if we had the kind of platform I've been running my mouth about for years, we wouldn't need to suffer configuration- and rollout problems.
comment in response to
post
I now think "mutable" would have been a better choice. Otherwise, if it doesn't change state, it doesn't take any information from the data, right? A transformer is not an observer, but it might have a configuration, so some final state.