
Aix-Marseille University generates interest amid a US crackdown and calls for a ‘scientific refugee’ status

Programmer and sysadmin (DevOps?), wannabe polymath in tech, science and the mind. Neurodivergent, disabled, burned out, and close to throwing in the towel, but still liking ponies 🦄 and sometimes willing to discuss stuff.
Hm, good point. I generally go on feeling, from an English as an Nth Language point of view... and my subjective feeling is that "snuck" has more of a "participle" meaning, while "sneaked" has more of a "past tense" meaning.
According to AI Overview, there might also be some EN-US vs EN-GB at play:
"Snuck" is an irregular past tense: It's an alternative form that has gained widespread acceptance, especially in North American English.
"Snuck" is sometimes considered nonstandard in British English: While it's increasingly common in British English, it's still often seen as nonstandard in formal writing.
That would match the Wiktionary entry: https://en.m.wiktionary.org/wiki/sneaked
Well, technically... we have an example in modern Spain of an (almost) peaceful and willing transition without abdication:
Other than a failed coup attempt by a faction of the military who wanted to go back to the previous system, it was a reasonably peaceful transition from full dictatorship, to a "parliamentary monarchy".
It can be done, if people are willing.
(PS: an abdication came much later, because of some not fully transparent money deals and tax evasion schemes, leaving his son as the new King)
It reads like written by AI: some standard keywords, key phrases, an overall sentiment, and a few out-of-style words that sneaked in.
I doubt it's been fed text about "bergro", "parava", and "rortx", this looks like basic reasoning to me:
For the sake of completeness, this is qwen3:1.7b running on ollama on a smartphone. Its reasoning is more convoluted (and slow), yet the conclusion is the same:
If all bergro are rortx, and all parava are rortx, are all rortx parava?
Answer: No, not all rortx are parava. The premises do not establish a relationship between bergro and parava, so rortx could include elements from both groups.
"AI" has been a buzzword basically forever, it's a moving target of "simulates some human behavior". Every time it does that, we call it an "algorithm" and move the goalpost for "true AI".
I don't know if we'll ever get AGI, or even want to, or be able to tell if we get a post-AGI. Right now, "AI" stands for something between LLMs, and Agents with an LLM core. Agents benefit from MCP, so that's good for AI Agents.
We can offload some basic reasoning tasks to an LLM Agent, MCP connectors allow them to interact with other services, even other agents. A lot of knowledge is locked in the deep web, and in corporate knowledge bases. The way to access those safely, will be through agents deciding which knowledge to reveal. MCP is aiming to become the new web protocol for "AI"s, no less no more.
Some careless people will get burned, the rest will be fine.
I feel like a better solution is to get an AI SO. Shape them into whatever you like, don't forget it's still an AI, and get whatever comfort you need in the moment.
You can even have several at once.
I want an eye replacement with a few extra opsins, and some EM sense like what birds have. Eternal life could be nice too, would make a lot of people reconsider their long term plans.
The only way for tolerance to exist, is to not tolerate intolerance.
Necessary reminder:
The connectors are still optional.
Haphazard code is not a new thing. Some statistics claim that almost 50% of "vibe coded" websites have security flaws. It's not much different from the old "12345" password, or the "qwerty" one (not naming names, but have known people using it on government infrastructure), or the "who'd want to hack us?" attitude.
MCP is the right step forward, nothing wrong with it on itself.
People disregarding basic security practices... will suffer, as always... and I don't really see anything wrong with that either. Too bad for those forced to rely on them, but that's a legislative and regulatory issue, vote accordingly.
I would still be extremely hesitant of enabling any MCP connector on non-local model instances. People need to push harder for local and on-prem AI, it's the only sane way forward.
What a pity. The first book, and film, were great, mixing dark, fun, and morals. Somewhat in a Grimm Brothers way, somewhat in a magical way. How far has JKR fallen since 🙍
One of the worst possible examples ever: Klarna is a payment processor, people don't call their bank to get the same answer the system is already giving them, they call to negotiate something about their money. AIs are at a troubleshooting level, at best some very basic negotiation, nowhere near dealing with people actually concerned about their money... much less in 2023.
Seems like Klarna fell hook, line, and sinker for the hype. Tough luck, need to know the limits.
Randomly obfuscated database: you don't get exactly the same data, and most of the data is lost, but sometimes can get something similar to the data, if you manage to stumble upon the right prompt.
Yeah, I don't think I like llamafile, reusing some weights between models, and smaller updates, sounds like a better idea.
What I'd like to see is a unified WebNN support, for CPU, GPU, and NPU: WebNN Overview
(Not to pull rank, but my mail profile can be tracked to Netscape Navigator, across multiple OSs 😁)
It's going to be funnier: imagine throwing in tons of data at an LLM, most of the data will get abstracted and grouped, most will be extractable indirectly, some will be extractable verbatim... and any piece of it might be a hallucination, no guarantees! 😅.
Courts will have a field day with that.
That's why AI companies have been giving out generic chatbots for free, but charge for training domain-specific ones. People paying for using the generic ones, is just the tip of the iceberg.
The future is going to be local or on-prem LLMs, fine tuned on domain knowledge, most likely multiple ones per business/user. It is estimated that businesses are holding orders of magnitude more knowledge, than what has been available for AI training. Will also be interesting to see what kind of exfiltration becomes possible, when one of those internal LLMs gets leaked.
Is Disney no longer a "pal", or did it stop making money?
All of them. The moment they summarize results, it automatically filters out all the chaff. Doesn't mean what's left is necessarily true, just like publishing a paper doesn't mean it wasn't p-hacked, but all the boilerplate used for generating content and SEO, is gone.
Starting with Google's AI Overview, all the way to chatbots in "research" mode, or AI agents, they return the original "bulletpoint" that stuff was generated from.
Can you elaborate? It does match my personal experience, and I've been on both ends of the trash flinging.
A lot of people have been working tedious and repetitive "filler" jobs.
In the near future, AI-controlled robots are going to start replacing low skilled labor, then intermediate skilled ones.
"AI" has the meaning of machines replacing what used to require humans to perform. It's a moving goalpost: once one is achieved, we call it an "algorithm" and move to the next one, and again, and again.
Right now, LLMs are at the core of most AI, but AI has already moved past that, to "AI Agents", which is a fancy way of saying "a loop of an LLM and some other tools". There are already talks of moving past that too, the next goalpost.
Aix-Marseille University generates interest amid a US crackdown and calls for a ‘scientific refugee’ status
What they were offering – through a programme titled Safe Place for Science – was a sort of “scientific asylum”, offering three years of funding at their facility for about 20 researchers.
On Thursday the university said it had received 298 applications in a month, of which 242 were deemed eligible. The applicants hailed from institutions such as Johns Hopkins University, Nasa, Columbia, Yale and Stanford, it said in a statement.
Most of the applications were sent using encrypted messaging, the university’s president, Eric Berton, wrote in the French newspaper Libération.
State tells employees to report on one another for ‘anti-Christian bias’
The Trump administration has ordered State Department employees to report on any instances of coworkers displaying “anti-Christian bias” as part of its effort to implement a sweeping new executive order on supporting employees of Christian faith working in the federal government.
The cable was sent out to embassies around the world under Secretary of State Marco Rubio’s name. The instructions also were released in a department-wide notice.
The cable encourages State Department employees to report on one another through a tip form that can be anonymous. “Reports should be as detailed as possible, including names, dates, locations (e.g. post or domestic office where the incident occurred,” the cable reads.
“It’s very ‘Handmaid’s Tale'-esque,” said one State Department official, who was granted anonymity because the individual was not allowed to speak openly about internal department affairs.
New approach punishes AI companies that ignore “no crawl” directives.
Popular Chrome extensions hijacked by hackers in widespread cyberattack
Legitimate browser extensions were turned bad through malicious updates
a number of popular extensions that enable things like dark mode and adblocking in Google’s browser have been hijacked by hackers, putting 3.2 million Chrome users at risk.
While all of the extensions listed below have since been removed from the Chrome Web Store, you will still need to manually delete them if they’re currently installed in your browser
What Is the Third Term Project? Pro-Trump Group Bids to Change Constitution
The project wants more Republicans to support Representative Andy Ogles' push to allow Trump to run for office in 2028.
A Republican group is hoping to rally support to change the Constitution to allow President Donald Trump to seek a third term.
The 22nd Amendment to the Constitution was ratified in 1951 following the presidency of Franklin D. Roosevelt, who was elected to four terms between 1933 and 1945. The two-term limit for presidents was introduced by Congress to prevent potential abuses of power.
Trump has already sought to test the limits of executive power.
The official White House social media accounts on X, Instagram and Facebook soon quoted his post, all sharing a fake magazine cover depicting an illustration of Trump smiling in a suit — and wearing a bejeweled golden crown.
Netanyahu gifted Trump a golden pager during their meeting in Washington
"Press with both hands"
...just when you thought this timeline couldn't get much weirder.
US Immigration and Customs Enforcement is in discussions with Phoenix-based Willscot about a lease, according to people familiar with the matter.
US Immigration and Customs Enforcement is in discussions with Phoenix-based Willscot about leasing the company’s mobile structures to house undocumented detainees, the people said. Willscot’s products are commonly used as construction-site storage and office space.
This AI Paper Unveils the Future of MultiModal Large Language Models (MM-LLMs) – Understanding Their Evolution, Capabilities, and Impact on AI Research
Brace for impact.
Israeli troops briefly enter Gaza as wider ground incursion looms
The Israeli military says its troops and tanks briefly entered northern Gaza overnight into Thursday, hitting several militant targets as a wider ground incursion looms after more than two weeks of heavy air raids
Israeli troops and tanks launched a brief ground raid into northern Gaza overnight into Thursday, the military said, striking several militant targets in order to “prepare the battlefield” ahead of a widely expected ground invasion
Deleted posts
It's unnerving to find an interesting post, with an interesting conversation, only to see it deleted (not even mod removed) with hanging replies in the inbox and no way to reply back.
Is there any feature that would allow continuing those conversations? Other than direct messages, which get "black holed" (no way to see own replies). Could these conversations be somehow continued, either recovered in Lemmy, or maybe via Mastodon?
Google Gmail continuously nagging to enable Enhanced Safe Browsing
The difference between the two security features is that Safe Browsing will compare a visited site to a locally stored list of domains, compared to Enhanced Safe Browser, which will check if a site is malicious in real-time against Google's cloud services.
While it may seem like Enhanced Safe Browsing is the better way to go, there is a slight trade-off in privacy, as Chrome and Gmail will share URLs with Google to check if they are malicious and temporarily associate this information with your signed-in Google account.
Another room-temperature superconductor
The room-temperature superconducting space has suddenly become white hot with new papers and patent applications.
This time, straight from a patent granted to a blockchain company, with no accompanying paper or proof.
Edit: after reviewing the patent, and as pointed out by @floofloof@lemmy.ca, this is an incredible amount of BS. The patent's initial date is Feb 2020, issue date Dec 2021. It has no proof, because it claims to speculatively apply a possible theory by someone else, onto how to make a flexible Type II semiconductor out of a Type I semiconductor, in case this ever happens to be possible with that theory. Basically a patent troll waiting to see if someone happens to make possible the elements they've used in the patent, then jump in and claim an application.
Honestly, didn't know speculative patents like this were possible.