barrkel

13 ms·

211.
by barrkel 10mo ago
You shouldn't generally use PKs as public identifiers, least of all UUIDs, which are pretty user hostile.
212.
by barrkel 10mo ago
UUID v7 doesn't squeeze creation date in. If you treat it as anything other than a random sequence in your applications, you're just wrong.
213.
by barrkel 10mo ago
Uuid v7 just has a bias in its generation; it isn't carrying information. You're not going to try and extract a timestamp from a uuid. Random vs time biased uuids are not a decision to shave off ms that you will regret. Most likel
214.
by barrkel 10mo ago
It was also used by Delphi in 90s.
215.
by barrkel 10mo ago
The model giving it text to speak would have to annotate the text in order for the TTS to add the affect. The TTS wouldn't "remember" such instructions from a speech to text stage previously.
216.
by barrkel 10mo ago
Incentives rule everything. For the Romans, winning wars was the main source of elite prestige. So the Empire had to expand to accommodate winning more wars. Today, the stock market and material wealth dominates. If elite dominance of the m
217.
by barrkel 10mo ago
Look here: https://vulert.com/vuln-db/CVE-2025-48633 It has to do with setting the device owner, and gaining those powers; enabling / disabling apps, remote wipe, etc.. It's a local privilege escalation attac
218.
by barrkel 10mo ago
How do you e.g. validate that a database product works with all the different cloud databases? Every time you change up SQL generation you're going to want to make sure the SQL parses and evaluates as expected on all supported platform
219.
by barrkel 10mo ago
Transitive dependencies?
220.
by barrkel 10mo ago
You specifying the top level hash doesn't do anything to pin transitive dependencies, and as the article points out, transitive dependencies - especially dependencies common to a lot of actions - would be the juciest target for a suppl
221.
by barrkel 10mo ago
No, it's more like they sell to a bunch of verticals and they have checklist-orientated market segmentation where they can eke out a little extra profit selling a specific line to a specific industry because they need a specific doodad
222.
by barrkel 10mo ago
Of course different strokes for different folks. My favorite laptop at this point is a terminal for my home workstation which is far more powerful than anything mobile. That means I prioritize decent graphical performance, OLED screen, long
223.
by barrkel 11mo ago
A little over 20k CHF. I delegated building (and ensuring the parts work together and testing) to Dalco, a Swiss company that specializes in academic compute clusters and the like. The price was competitive with building it myself. The CPU
224.
by barrkel 11mo ago
It's got that AI smell. The word choice. Antithesis ("not this, but that"). Persuasion words. There's meat in it, it's not pure slop, but it was definitely fed through the slop factory. And as some have mentioned, t
225.
by barrkel 11mo ago
I feel this vibe. It's part of the reason I invested in a monster home workstation - threadripper 9995wx, 768gb ecc ram, 96gb Blackwell pro. I expect it may easily be the last proper home PC I buy that is scaled in line with the scalin
226.
by barrkel 11mo ago
The non-convex side mirror almost got me into an accident the first rental car I drove in the US. I was expecting to see more of the road than I did.
227.
by barrkel 11mo ago
It's weird how this seems to be coming everywhere all at once. Is it just politician FOMO, when they see someone trying this somewhere they feel they need to try it too? It feels like a concerted push to deanonymize the web.
228.
by barrkel 11mo ago
I guess I lucked out. I bought a 768GB workstation (with 9995wx CPU and rtx 6000 Pro Blackwell GPU) in August. 96GB modules were better value than 128GB. That build would be a good bit pricier today looks like.
229.
by barrkel 11mo ago
If the observable behavior is bad Linux performance, it's a Linux problem. There's a saying in motorcycling: it's better to be alive than right. There's no upside in being correct if it leaves you worse off. There are wa
230.
by barrkel 11mo ago
The structure of what it wrote, and the banality of the point.
231.
by barrkel 11mo ago
This would speed things up since it looks like the bottleneck here is I/O.
232.
by barrkel 11mo ago
The problem with this is you don't have context on subgraphs of control flow which have already returned by the time of the error. I think our disagreement is less about error handling and more about logging policy. I favour a logging
233.
by barrkel 11mo ago
The information about what the code was doing at the time an exception is thrown is implicit in the stack trace; the control flow is evident from line numbers, and with half-decent tooling, you can get hyperlinks from stack traces to source
234.
by barrkel 11mo ago
I believe the formalisms can be constructed by something fuzzy. Humans are fuzzy; they create imperefect formalisms that work until they break, and then they're abandoned or adapted. I don't see how LLMs are significantly differen
235.
by barrkel 11mo ago
This is error-prone boilerplate that obscures the code, obscures the logs and is a known antipattern (log and throw) - which you're implementing manually, by hand, in the hope you never make a mistake. You shouldn't do manually wh
236.
by barrkel 11mo ago
So you want to gradually reduce the fidelity of the error message as it makes its way up the stack. That means that the top level handler can at best log a vague message. That in turn means you must log along the way where you have more p
237.
by barrkel 11mo ago
I guess you abandoned PC gaming some time in the early 2000s?
238.
by barrkel 11mo ago
Don't forget, failure modes pierce abstraction boundaries. An abstraction that fully specifies failure modes leaks its implementation. This is why I think checked exceptions are a dreadful idea; that, and the misguided idea that you sh
239.
by barrkel 11mo ago
You probably do want that exception to bubble up, actually. You probably don't want to catch it immediately after open. Because you need to communicate a failure mode to your caller, and what are you going to do then? Throw another exc
240.
by barrkel 11mo ago
Knowledge models, like ontologies, always seem suspect to me; like they promise a schema for crisp binary facts, when the world is full of probabilistic and fuzzy information loosely categorized by fallible humans based on an ever slowly sh

More ›