gojomo

27 ms·

331.
by gojomo 4y ago
Under what conditions do you suggest "Google itself also has access to all your 2FA secrets"? (Without cloud backup, & without the installation of a malicious version of 'Google Authenticator', how would they – espec
332.
by gojomo 4y ago
Yep. Reads a lot like the public cope & FUD of a player behind & not ready to admit it. It's a bit like all the companies caught flat-footed by the iPhone, pooh-poohing its prospects in public – even if furiously trying to cat
333.
by gojomo 4y ago
It took over 30 years to go from, say, the Model T (1908) to the 1st automatic transmission in a mass-produced passenger car (1939). The 1st "DARPA Grand Challenge" was in 2004 – when not a single entrant made it further than 8 mi
334.
by gojomo 4y ago
There have been self-driving busses that use such painted guides on fixed routes; here's a story where Las Vegas was testing such a system using the tech of 22 years ago: https://www.nytimes.com/2001/07/26&#x2
335.
by gojomo 4y ago
That seems like an experimentally-testable prediction - that attempts to "clone ChatGPT's behavior (by using its output)" will necessarily "get the worst of both worlds: weird output coming from its RLHF step AND halluci
336.
by gojomo 4y ago
Yep. And even without the perceptual/disassociative/hallucinogenic aspects of heavy use, daily nitrous exposure, without long periods away, apparently interferes with Vitamin B12 metabolism enough to cause nervous system damage. E
337.
by gojomo 4y ago
Even before Hsieh left Zappos, anecdotes I heard from the fringes of his circle, including people who had worked at Zappos, made it clear he operated in a very heavy-drinking culture. For example, both media & employee reports mention
338.
by gojomo 4y ago
If the reporter hasn't received a list of the things tested-for-and-negative, from a representative of the competent institution that did the testing, that's definitionally 'hearsay' in a legal context. Even in a lay c
339.
by gojomo 4y ago
Agreed: the dispositive factors are: "Is this true?" and "Could this truth help others?" Given the other reports of Hsieh's erratic behavior, including with regard to health practices & psychoactive drugs, &
340.
by gojomo 4y ago
Totally agreed, but still, finder's fees sometimes work that way – because the actual amount truly paid is considered the least-fakable proxy-signal of value delivered. Similarly, utility-style government regulation often caps profits
341.
by gojomo 4y ago
That's useful context - but both those details in the article seem like indirect hearsay, from people who don't necessarily have the whole story. For example, the drug info isn't from an autopsy, but what some unnamed UC Davi
342.
by gojomo 4y ago
Both wealth and drugs can cause small personality quirks, & slight mental instability, to grow in unchecked proportions. Synergistically, especially, they can separate people from familiar moderating social situations, and introduce the
343.
by gojomo 4y ago
Oddly, it's not that different from other commission/finder's-fee systems. (It's also not totally different from the "commission" many governments collect via sales taxes: "everything happening within ou
344.
by gojomo 4y ago
You wouldn't want to spread "this information" even if it happened to be true , and such truth might help others either navigate their grief, or recognize & head-off similar delusions in the future? FWIW, the detail abou
345.
by gojomo 4y ago
There's only the testimony of the financial manager (Tony Lee) to imply that Tony Hsieh's brother Andy was "seeking to exploit his brother rather than protect him". So I think it remains plausible that Andy was , in tru
346.
by gojomo 4y ago
That's an interesting angle, and barring explicit clarification from OpenAI, it'd be possible to see 'ChatGPT' either as an 'OpenAI API' (in a very general sense), or as not the same thing they mean by &#x2
347.
by gojomo 4y ago
Indeed, & an artificial uniformity of LLMs, and thought, if everyone is cribbing each others' outputs could be a concern. But before its concern about a monoculture, the article 1st points out that mere prediction-training (on an
348.
by gojomo 4y ago
A watermark is absolutely possible - see for example some of the work Scott Aaronson has mentioned doing for OpenAI. But: very fragile, especially if people are specifically trying to hide their GPT use, or have access to the watermarking a
349.
by gojomo 4y ago
Yes, the inner link to the Yoav Go writeup is the gem here, concisely explaining the benefits of RLHF-scoring over mere prediction. Though, as Go speculates, it's likely possible to reduce even further the HF ("human feedback&quot
350.
by gojomo 4y ago
While I'm eager for people to be able to use leading-edge LLMs without associating their prompts with the real identity, this offering seems untenable. It's at the mercy of OpenAI to throttle or disable the underlying access. It&#
351.
by gojomo 4y ago
There's no report from any investors who suggest they were confused about IP/original-work in any way. There is an assertion to the contrary in the article: "Stability told Sifted that the investors that backed the company
352.
by gojomo 4y ago
Mostquito-larva-eating fish have often been intentionally deployed for mosquito control, though the use of non-native species may have sometimes backfired: https://en.wikipedia.org/wiki/Mosquitofish#Environmental_imp...
353.
by gojomo 4y ago
Also doesn't work in (Chromium-based) Brave. Might be Google product-tying its YouTube dominance to Chrome dominance.
354.
by gojomo 4y ago
Thy've been proactive in deleting chat logs that they were ordered by courts to keep for pending litigation.
355.
by gojomo 4y ago
Sure, a reorg'll fix things.
356.
by gojomo 4y ago
That sounds way high, even when the antigen tests were new, if the 'true' condition is considered PCR-positivity. Rapid antigens miss tons of PCR-positives, but that's not as bad as it sounds because plenty of those PCR pos
357.
by gojomo 4y ago
Aren't allergic reactions to nanotech detritus in the atmosphere a bit of background detail in Stephenson's 'Diamond Age' (1995)?
358.
by gojomo 4y ago
Same here.
359.
by gojomo 4y ago
Seems plausible. Suffered a SIM-hijack attack via T-Mobile a few years ago. Set a giant extra arbitrary password for account changes after that - but they essentially don't ask for it. Fairly regularly they show notices of breaches via
360.
by gojomo 4y ago
...and this tiny fragment from the self-report of one aggrieved party is quite possibly far less hostile than whatever full comments/history actually triggered a "100th time" subreddit ban.

More ›