In my opinion, the only immoral part of this is the “million dollars in profit” on the house in Palo Alto that he left which the author himself called out as ridiculous. This unearned increment is thanks to Palo Alto’s exclusionary zoning and Proposition 13’s low property taxes (around 1.25% at acquisition, but much lower for long-term investors). ChatGPT says that the author paid $11k/year and the new owner now pays $28k/year for the Palo Alto house now assessed at $2.3M.

Abuses in violation of policy are a problem like in any organization that can be reduced through auditing, discipline, etc. And it seems that the ones that have been caught (Jarmarus Brown of Orange City, FL; Asad Zahir of Shively, KY; Emily Pacheco of New Bedford, MA) have been disciplined and/or charged.

Insufficient security of devices is a problem but does not have to do with my comment about the intended uses.

Officers have to be trained about OCR errors, just like we have to know when to trust an LLM. The SFPD policy I linked says officers shall “Visually verify the alphanumeric characters on the plate” before stopping a car identified by ALPR. And like LLM hallucinations OCR errors will go down over time.

Locate stolen, wanted, and or other vehicles that are the subject of investigation

To apprehend wanted persons subject to arrest warrants or who are otherwise lawfully sought by law enforcement.

To locate victims, witnesses, suspects, missing children, adults, and/or elderly individuals, including in response to Amber Alerts and Silver Alerts and others associated with a law enforcement investigation.

To assist with criminal investigations initiated by local, state and regional public safety departments by identifying vehicles associated with targets of criminal investigations.

Counter-terrorism: Identify potential threats to critical infrastructure sites.

For other law enforcement purposes as authorized by law: Investigations of major crimes.

Which of these uses do people not like? Or which other policies are people concerned with in other jurisdictions?

That’s not my argument at all. We do have federal antitrust law. My question was what does the ordinance do that differs from antitrust. The answer is that the ordinance rewards bounty hunters chasing the same federal case, and also bans non-“competitively sensitive” datasets.

There’s proportional punishment based on harm, and then there are abusive booby-traps. If a landlord uses a nonpublic dataset (e.g., a vendor’s model based on year-old data that would not harm competition under the antitrust settlement), then a tenant and nonprofit can each sue for civil penalties plus attorney fees even though there is no harm.

1. It goes beyond antitrust guidelines by banning rent prediction using any “non-public competitor data”, which is stricter than the former antitrust guidelines which prohibited sharing “competitively sensitive variables” (https://www.ftc.gov/sites/default/files/attachments/dealings...). To the extent that it differs from banning competitively sensitive data, it is just banning the use of data to make rentals more efficient. If landlords are pricing accurately but not monopolistically, this should reduce turnover, reduce vacancy, and reduce the occasions where a listing gets a crowd of applicants, which is good for both landlord and tenant. It’s bad to try to ban accurate data.

2. It provides a private right of action for tenants and nonprofits to sue, just like Proposition 65 (Known by the State of California to cause cancer) incentivizes bounty hunters to sue. The point seems to be to punish landlords more than it is to establish fair rules.

And it seems that these lawsuits are just piggybacking on the DOJ settlement by punishing anyone who uses RealPage as soon as the ordinance becomes effective in 30 days while RealPage was already working to settle federally. The complaints (e.g. the SF one https://www.courtlistener.com/docket/73573242/gomez-v-greyst...) don’t have any new violations; just companies who are already settling. I don’t think that’s justice for a city to frontrun the federal settlement with new violations for the same program.

This is hysterical. No, format shifting old unwanted books is not unethical. The books still exist, in an internal digital library. If copyright law were to change to allow sharing orphan works some day, Anthropic could share them. But under current law, the books are preserved digitally and used for transformative uses that all Claude users benefit from.

No, destroying collectible books would be a shame, not just any rare worthless books. But these are not collectible. The article tried to dance around it by saying maybe some books have a sentimental value to someone somewhere. But that doesn’t mean any library or collector wants it. Don’t fall for manufactured outrage!

Edit: Here’s an example of an extremely rare book. My great great grandfather published a book of sermons around 1920. That book has zero value to anyone other than my dad. Would I be outraged if it ended up at someone’s estate sale, then a used bookstore, and then an LLM consumed it to learn to read? No; I would have expected it to have been discarded by humans before the LLM even got to it. Most of what we leave behind is discarded.

Legally, this case is about the terminology of a deed that was sloppily made “in trust… for parkland”, and standing to sue. It’s currently on appeal. (Trial court: https://judicialrecords.wilco.org/PublicAccess/CaseDetail.as.... Appeal, 15th Court of Appeals: https://search.txcourts.gov/Case.aspx?cn=03-25-00831-CV&coa=.... Appeal, 15th Court of Appeals: https://search.txcourts.gov/Case.aspx?cn=15-25-00202-CV&coa=...).

It’s a shame that the deed was poorly granted. Perhaps it would have been better for everyone if it had been held privately and taxed as such.

And as far as newsworthiness is concerned, the actual deed lost its restriction in 2003, and the city transferred it to the Economic Development Corporation in 2008. It could have then been sold to any industrial purpose. This is not really a national story about datacenters. It is the story about a 2008 sale for industrial purposes. The fact that this is recirculated as a datacenter story is meant to poison the public on any mention of the word datacenter. I think this trend of biased news is intentional.

According to opus 4.8, here’s the chain of title for the 87.797‑acre Taylor tract:

1. 7/7/1999 — Bonnibel Bland Cromwell & family → Texas Parks & Recreation Foundation (deed carries the "held in trust… for parkland" restriction). Recorded #199947198. Consideration: $10 (charitable donation).

2. 10/7/2003 — Texas Parks & Recreation Foundation → Williamson County Park Foundation, Inc. Recorded #2003100356. Consideration: $10 (nominal/recited).

3. 11/20/2003 — Williamson County Park Foundation → City of Taylor. Consideration: $10 (nominal/recited).

4. 11/12/2008 — City of Taylor → Taylor Economic Development Corporation. Recorded #2008084718. Consideration: $15,000 cash + a land swap — the EDC also conveyed two tracts back to the City “by exchange” (a ~22.708‑ac tract in the Samuel Pharass Survey + a 16.658‑ac tract in the Coursey Survey); no dollar value stated for the swapped tracts.

5. 11/19/2024 — (Plat Map Recording Sheet; not a transfer)

6. 4/11/2025 — Taylor Economic Development Corporation → NCP Travis TPP Project, LLC (Blueprint Data Centers). Consideration on the deed: "Cash and other good and valuable consideration" — no figure stated; reported at ~$10 million in the press.

And I am concerned that the purpose of slanted anti-datacenter coverage by the likes of 404media.co and perfectunion.us is to inspire memetic NIMBYism that has and will cause tremendous damage to the US.

And I am concerned that the purpose of slanted anti-datacenter coverage by the likes of 404media.co and perfectunion.us is to inspire memetic NIMBYism that has and will cause tremendous damage to the US.

I like the idea of the more ergonomic, faster api in new-stream with no buffering except at Stream.push(). NodeJS and web streams put infinitely expandable queues at every ReadableStream and WritableStream so that you can synchronously res.write(chunk) as much as you want with abandon. This API basically forces you to use generators that yield instead of synchronously writing chunks.