I like the idea, but it seems hard to make a law that it is illegal to go broke. I think the law should be that the server code is distributable and licenced to the users in the event of service failure.
Posts by John69
61 publicly visible posts • joined 9 Mar 2022
$800 'AI' robot for kids bites the dust along with its maker
The only thing worse than being fired is scammers fooling you into thinking you're fired
Will passkeys ever replace passwords? Can they?
It seems most of the complains are based on the system being used. I am no expert, and I gave Passkeys a quick test. Using ubuntu and KeePassXC I got a fully backed up and transferable Passkey login to Github in only a few minutes with no hardware or external service required and no confusing multiple systems asking me to do the job.
The real problem is the requirement to have a password backup, but this is not inherent to the Passkey technology but the implementations by the servers.
Re: Just promote the sensible use of passwords, teaching it in schools.
"Some can't cope with anything more complicated than passwords." Passwords are the most difficult to use. The user has to determine if the entity they are talking to is the same as who they talked to last time. HTTPS sort of provides this via a third party, but it is not easy. A system that does not require one to do this should be easier. The dark web manages it with PGP, how can the rest of the world make it so hard?
WordPress bans WP Engine from sponsoring or participating in user groups
Smart homes may be a bright idea, just not for the dim bulbs who live in 'em
FTC urged to stop tech makers downgrading devices after you've bought them
Re: Car thing
There is spotify car thing, where some people may not be able to listen to music, but what about the actual car thing? [Fisker owners](https://www.reuters.com/business/autos-transportation/ev-startup-fisker-files-bankruptcy-2024-06-18/) are finding out about that, and imagine how the UK would have handled it if Rover going bust had bricked all Rover cars? It seems this should be top of the list when it comes to legislation about digital resiliancy.
Brit teachers are getting AI sidekicks to help with marking and lesson plans
The cybersecurity QA trifecta of fail that may burn down the world
Why expect them to?
The idea that we should give capitalists all the power, and expect them to act in the best interests of society is completely naive.
The only answer is for us all to run the algorithms to block this stuff, eg. by training our own AIs from Hugging Face on what we want want blocked. Or we could take the power from the capitalists in a more direct way.
Privacy expert put away for 9 years after 'grotesque' cyberstalking campaign
War on Texas law requiring ID to savor smut online heads to Supreme Court
How two brothers allegedly swiped $25M in a 12-second Ethereum heist
I could be missing something here, but could not the crypto woo be stripped out of this story leaving "The brothers made an order for $25m of things, the traders bought the things for $25m from the brothers and they never honored the original order". This does not on the face of it seem the most novel model of financial fraud, and it would seem that procedures to prevent have developed over the millenia. Have the crypto traders of today forgotten what the Phoenicians knew of commerce?
Microsoft 365 Copilot 'generally available' – if you can afford 300 seats
Can this be done in open source?
It seems the components are there. With all the models on HuggingFace, libreoffice and the noncommercial text and data mining exception to copyright this shoud be doable and it shoudl be able to eclipse the non-comercial models is they are restricted in what they can use to train.
Google says public data is fair game for training its AIs
Scientists think they may have cracked life support for Martian occupation
Deepfakes being used in 'sextortion' scams, FBI warns
Is this any different from photoshop?
There was a time when if you saw a photo then what was depicted must have happened. Then photoshop came along and there was a short time when some people could be fooled by a photoshopped image. Then everyone learned and photoshop became an everyday tool.
I see no reason to think deep fakes will be any different. Once we learn we cannot trust video this will be no different to photoshoping someone head into a sex scene.
I do not believe there is much future for freely available AI image detecting software. The nature of generative adversarial networks means people will just incorporate whatever detection tool in the learning algorithm.
GitHub, Microsoft, OpenAI fail to wriggle out of Copilot copyright lawsuit
Re: Not at all
What CoPilot is learning, and this is the same for all LLMs, is what the most likely next word is given the preceding words. How exactly the output relates to the input in a legal sense is something the courts will decide. How similar that process is to human learning is something we shall all have to figure out.
Online Safety Bill age checks? We won't do 'em, says Wikipedia
British industry calls for regulation of autonomous vehicles
Re: We do not want "British" regulations
Historically national regulations come before international ones. Obviously the right international regulations will be better than the right national regulations, but that is no reason not to implement the national regulations before the technology hits the street.
Cyber-snoops broke into US military contractor, stole data, hid for months
Rather than take the L, Amazon sues state that dared criticize warehouse safety
DoJ ‘very disappointed’ with probation sentence for Capital One hacker Paige Thompson
Tesla Megapack battery ignites at substation after less than 6 months
US accident investigators want alcohol breathalyzers in all new vehicles
China discovers unknown mineral on the moon, names it Changesite-(Y)
LabMD gets another shot at defamation claim against 'extortionate' infosec biz
Surely if Tiversa was able at access this document then LabMD was as guilty as if the document had been put on bittorrent? If their security was so bad that Tiversa could access the file it was only luck that hackers did not make it available. Like claiming drink driving is fine as long as you do not kill anyone?
Woman forced to sell 4-bed house after crypto exchange wrongly refunded $7.2m
Tesla faces Autopilot lawsuit alleging phantom braking
California lawmakers approve online privacy law for kids. Which may turn websites into identity checkpoints
Amazon has repackaged surveillance capitalism as reality TV
Airbnb turns its anti-partying tech on American lodgers
> Airbnb has said that long-term stays "of 28 days or more" remain its fastest-growing category
Is it only me for whom this is the most worrying thing? Does this not indicate that AirBnB is now an accommodation provider rather than a holiday provider? When I stayed at an AirBnB in NY everyone else there was living there not on holiday.
Excel @ mentions approach general availability on the desktop
Re: Gerrof my lawn!
Quite. What does "assigning of tasks using the @ mention" actually mean? Are people using excel as a task tracking system? Is this a task tracking system for people who's tasks involve excel? Is this an IM client?
I generally thing that if you are using excel for anything that matters you are using the wrong tool, but I am quite sure that using it as a task tracker is wrong.
Facebook hands over chats to cops in abortion case
> a quick search shows me stats of 37% survival rate for babies born at 23 weeks
The number is very context dependent. For comparison, the WHO says 90% to those born under 28 weeks die within the first few days of life in poorer countries. What the number is for a poor woman in a country with one of the highest maternal mortality in west is certainly a question but I would be surprised if it is that high.
Charter told to pay $7.3b in damages after cable installer murders grandmother
Meta accuses data scrapers of taking more than their share
FBI and MI5 bosses: China cheats and steals at massive scale
Tech world may face huge fines if it doesn't scrub CSAM from encrypted chats
If they can do why do they not tell us how?
"We, and other child safety and tech experts, believe that it is possible to implement end-to-end encryption in a way that preserves users' right to privacy, while ensuring children remain safe online." They believe this, but refuse to say what leads them to believe this. Open source implementations of E2E encryption have been around for ages, if it was possible then they could easily demonstrate it.
Taiwan creates new challenge for tech industry: stern content regulation laws
California's attempt to protect kids online could end adults' internet anonymity
FBI warning: Crooks are using deepfake videos in interviews for remote gigs
What role does the deep fake play?
It is not clear exactly what the deep fake is doing here. Is it just about the ethnicity of the applicant, in that the company would not employ someone who looks Korean or whatever? Or is this pretending to be actual people who put their real identifiable photograph on the web (seriously, who does that in this day and age of scammers scraping the web for photos)? If you are giving out security information just because an applicant looks a bit like a picture on the web deep fakes are the least of your problems.
It seems like if this is a security hole then it has been created by the companies for no good reason.
Totaled Tesla goes up in flames three weeks after crash
Re: Deja vu again
I think the best answer is removable batteries. As well as this issue, it makes the problem of degrading batteries less, makes it quick to change rather than slow to recharge, allows faster adoption as the users do not have to buy the batteries, and makes closed loop recycling easier.
AI's most convincing conversations are not what they seem
Re: The real issue
That does not sound like a definition of sentience. If one made a machine that was functionally identical to the human brain except "you call a function with words and it returns an answer and then everything stops" would that make it non-sentient? We need a usable definition before we make a machine that fits it.
The real issue
It seems this, along with most commentary, is missing the main issue. Few if anyone thinks LaMDA is really a sentient being worthy of rights. The issue is that we do not have the tools to determine if it is or not, so we just come up with unfalsifiable statements that it is not, or what it means to be (both in this and the linked "expert" article). If we cannot distinguish LaMDA from a child who has only been exposed to the trillions of lines of text that LaMDA has been trained on, then we should spend the research money to making sure we do have the tools before we actually build an AI that may be a sentient being worthy of rights.
Businesses brace for quantum computing disruption by end of decade
Quantum computing startup probed in report, securities suit
Re: Quantum startups
Not that I like the the people who do this, if we vote for a system that allows people to make money this way can we really complain that some people chose to do so? Should we not "hate" the politicians who let them to it, or the voters who put the politicians in place? But that may get a little closer to home.
Threat of cross-border data tariffs looms over WTO
Re: "taxing e-commerce the same way that [..] physical goods traded internationally"
> Competitiveness is driven by competition not protectionism.
Historically all major economies have used protectionism to become competitive. They generally only go free market fundamentalist once they are globally competitive.