CEO, PGX Inc. / @galloprehistorico.bsky.social / @cep@fosstodon.org

no highly esteemed deed is…
Ending my involvement with Twitter. See everyone at Blusky: galloprehistorico.bsky.socia…
4
308
One thing I hope every developer takes from the xz situation is that, if someone attempts to shame and badger you into accepting patches on any project, the answer should be, "If you are having trouble finding the 'fork' button, I can send you a screenshot."
1
1
439
PostgreSQL privileges can surprise you. thebuild.com/blog/2024/03/25…
1
2
341
Christophe Pettus retweeted
Christophe Pettus: Checking Your Privileges postgr.es/p/6sb
5
20
2,858
Christophe Pettus retweeted
Christophe Pettus with an important message for anyone interested in PGX's services! @Xof @PostgreSQL @nordicpgday #postgres
1
6
382
"AI" customer service chatbots are just incredibly, horribly, laughably bad. They are like someone is conducting a sociology experiment in how much it takes to piss customers off. "I need help correcting a W-2." "Got it! Here are some topics about waffles!"
1
220
I wish fewer conversations with clients were of the form: Client: "We are driving nails with a screwdriver, and it is not working well." Us: "Use a hammer?" Client: "It's too risky to introduce a new tool into our system. How can we make the screwdriver work better?"
2
1
9
493
I think that Kissinger waited until this moment to die so he could kick off at a time when it could cause the largest number of people to scream at each other over the correct interpretation of his legacy. (Which is silly, because his legacy is that he was a war criminal.)
2
171
I am hereby defining a new problem in programming languages, which is: Syntax creep. Python and C++ display this perfectly. The language starts with a (relatively) simple and direct syntax, and slowly adds more and more appurtenances until it becomes totally unreadable.
2
276
"Having reliable audio on a laptop is bad, actually."
2
307
Christophe Pettus retweeted
I still haven't heard a good answer to this question, on or off the podcast. AI researchers often tell me, "Don't worry bout it, scale solves this." But what is the rebuttal to someone who argues that this indicates a fundamental limitation?
702
667
7,749
3,243,176
random_page_cost is the brown M&Ms of #PostgreSQL database tuning. If you look at the configuration and see that random_page_cost = 4, look through the rest of the configuration *very* carefully. npr.org/sections/therecord/2…
3
7
437
Christophe Pettus retweeted
Amen to that! "Lorenic said this flood of junk CVE 'is literally going to DOS triage teams, the NVD itself, and open source maintainers stuck dealing with the fallout. Stop doing this!' He's correct." opensourcewatch.beehiiv.com/…
1
4
16
1,490
Dear App Developers: Just so you know, the very first time I get a marketing notification on my phone, your app gets notifications turned off and never turned on again. Provide a clear way to opt-in, or don't use them for that.
1
7
249
Many (most?) cities already have zoning laws that, if interpreted strictly, prohibit people from working from home. This is not implausible. nitter.net/LeavittAlone/status/16…
229
Dear @Zoom: I would like you to state, unequivocally, if Zoom has the right to use user recordings as input to AI training, or not. Do you have that legal right, under the TOS as it stands, right now? Please answer without equivocation. 1/n
1
2
248
My reading of the ToS that Zoom does claim that right, right now, and the user cannot take that right away from Zoom. Zoom might "promise" that it won't do so if the user checks some box on the UI, but that does not take away Zoom's legal right to do so.
1
137
Further, the ToS and the blog post are both unclear about what happens if one user has taken action to prevent Zoom from using their data as input to AI training, but other parties on the call have not.
123