gojomo

14 ms·

151.
by gojomo 2y ago
Was there ever an unsweetened Pepsi?! No sugar or other sweeteners?
152.
by gojomo 2y ago
See also: 'broken windows theory': https://en.wikipedia.org/wiki/Broken_windows_theory Though this concept isn't as pat and reliable as in its most-simplistic formulations – you can't fix all crime
153.
by gojomo 2y ago
I like that your deep dive into whether this story might be true is the polar opposite of the "too good to check" impulse in journalism & social media, which instead tries to squeeze some attention & entertainment out of a
154.
by gojomo 2y ago
Interesting analogies for LLMs! ( https://en.wikipedia.org/wiki/W%5EX & https://en.wikipedia.org/wiki/Return-oriented_programming )
155.
by gojomo 2y ago
That's my read of how IA & publishers summarize the agreement. The IA has usually focused on things not available elsewhere, like out-of-print works – even the "National Emergency Library" was aiming to replace lending
156.
by gojomo 2y ago
There will always be lawsuits. But let's check back in 10 years to see if any of the specific lawsuits you predict materialize and have any effect on the IA. I doubt it; it's unlikely any litigous rightsholder with a strategic rea
157.
by gojomo 2y ago
Why isn't the usual "regulatory framework" around powerful new technology and products – common-law liability for harms, prohibitions against fraud & various kinds of trespass & damage to established rights, plus the
158.
by gojomo 2y ago
Some of the language in the Cornell page implies to me it's specific to "preservation" copies, not necessarily other format-shifts for lending. But the interactions of tradition, legal precedents, and different laws in differ
159.
by gojomo 2y ago
Roughly correct, but: the main issue turned out to be whether even the 1:1 lending was legal. That's what the initial ruling clearly denied IA, and contesting that is the focus of the IA's appeal. (The appeal also asks, in a small
160.
by gojomo 2y ago
The Internet Archive still thinks it has a legitimate fair-use, first-sale, & library rights basis for 1:1 controlled lending of ebooks for which physical copies are held. With the advice of competent counsel, IA clearly estimates its c
161.
by gojomo 2y ago
No, the publishers' lawsuit also took issue with the "NEL". It's just that the lower court's judgement didn't really focus much that issue, because they found the broader/simpler and more-foundational &quo
162.
by gojomo 2y ago
Everything you allege here is misinformed as to the current state of the lawsuit and stakes. You demonize Hachette et al (4 major publishers) as seeking to destroy a public good. In fact, they've already settled with the IA, as of last
163.
by gojomo 2y ago
From the headline to the details, this is a deeply misinformed take on the arguments, current case status, & possible outcomes. "Save itself"? Despite occasional prior histrionic kayfabe about "IA in existential danger&qu
164.
by gojomo 2y ago
That applies equally when each party is the filibuster-sized minority in the Senate. And: if Senate majorities really want to pass something, they can change the filibuster rules – and have, for some topics. Otherwise, the filibuster is mai
165.
by gojomo 2y ago
Perhaps because Democrats control half of Congress today, and the general trend of Congressional avoidance-of-clear-rulemaking has been the same even during those periods that Dems or Republicans control both chambers.
166.
by gojomo 2y ago
Lots of AI-progress-denialist cope is very ineffective! But the people demanding it want just-so stories that let them look away. Actual robust truth/effectiveness isn't what the market-for-cope is providing.
167.
by gojomo 2y ago
What are you trying to do, focusing on one parenthetical word, and not the bulk of the content? You could be a much better replier, without superficial word allergies.
168.
by gojomo 2y ago
This is useful as a "tiny paper" showing that one naive kind of self-training, in a smaller/older model (GPT-2), causes degradation. But it's likely those on a motivated search for evidence of LLM limitations ('cope
169.
by gojomo 2y ago
My read is that some former fans strongly disagree with, and are thus disappointed by, Tegmark's recent enthusiasm for "AI will kill us all" arguments, & advocacy of strong/intrusive policies against AI progress.
170.
by gojomo 3y ago
Yep! And as of the shutdown over 4 years ago, even if their control plane was insecure, they presumably had 119 working loudspeakers of the proper loudness/weatherproofing/etc.
171.
by gojomo 3y ago
These "sirens" actually need to be able to play spoken announcements - something they did easily since about 2005, before the fumbled 2019-2024 shutdown/upgrade. More history, through January 2022, via JWZ's blog: https
172.
by gojomo 3y ago
Remember when, less than 3 years ago, SF projected a $108M surplus for the next two years? Woulda been a nice time to clean up some of this technical debt! Or how about the SF Emergency Sirens, taken offline in late 2019 for a "2 yea
173.
by gojomo 3y ago
The hypercompetence of tech enthusiasts and the mind-numbing idiocy & dishonesty of politial governance don't mix, so cities have to spend tens of millions with parasitic contractors for what a few clever hackers could do as a side
174.
by gojomo 3y ago
Good question! Things that'd help: • a founders-duration copyright term – at most 28 years from date-of-creation • absolute statutory & precedential immunity for any copies made, or DRM-circumventing-technologies created/used,
175.
by gojomo 3y ago
Unless he was a guest lecturer, if the course was for credit, wouldn't his name appear on your official transcript?
176.
by gojomo 3y ago
You can tell this is authentically the industry insiders' own stunted understanding of the problem by the fact that the words 'copyright' and 'intellectual property' never appear as aspects of the challenge. Storage
177.
by gojomo 3y ago
Doesn't seem to have been as durably polarizing elsewhere, as far as I can tell. EG: even European countries that were more draconian earlier in the pandemic returned to relative normalcy, with less fear/debate around school reope
178.
by gojomo 3y ago
Unless you specifically infect a treatment group in an RCT, your 'recently had COVID' and 'no recently documented COVID' groups will vary in ways you're unlikely to be able to fully control-for, no matter how hard y
179.
by gojomo 3y ago
But when the gold rush is reaching a newly-discovered lode, at the end, there's real new gold! That they both eventually exhaust doesn't make them the same; one is fumes, the other involves a real expansion in the stock of usable
180.
Third Base (ternary notation for computing) (2001)(web.archive.org)
10 points by gojomo 3y ago|1 comments

More ›