#artificial intelligence interview process

[grplindia]

Why Choosing the Right Placement Agency in Lucknow Matters

Discover why selecting the right placement agency in Lucknow is essential for career success. Learn how expert guidance, exclusive opportuni

[frank-olivier]

Bayesian Active Exploration: A New Frontier in Artificial Intelligence

The field of artificial intelligence has seen tremendous growth and advancements in recent years, with various techniques and paradigms emerging to tackle complex problems in the field of machine learning, computer vision, and natural language processing. Two of these concepts that have attracted a lot of attention are active inference and Bayesian mechanics. Although both techniques have been researched separately, their synergy has the potential to revolutionize AI by creating more efficient, accurate, and effective systems.

Traditional machine learning algorithms rely on a passive approach, where the system receives data and updates its parameters without actively influencing the data collection process. However, this approach can have limitations, especially in complex and dynamic environments. Active interference, on the other hand, allows AI systems to take an active role in selecting the most informative data points or actions to collect more relevant information. In this way, active inference allows systems to adapt to changing environments, reducing the need for labeled data and improving the efficiency of learning and decision-making.

One of the first milestones in active inference was the development of the "query by committee" algorithm by Freund et al. in 1997. This algorithm used a committee of models to determine the most meaningful data points to capture, laying the foundation for future active learning techniques. Another important milestone was the introduction of "uncertainty sampling" by Lewis and Gale in 1994, which selected data points with the highest uncertainty or ambiguity to capture more information.

Bayesian mechanics, on the other hand, provides a probabilistic framework for reasoning and decision-making under uncertainty. By modeling complex systems using probability distributions, Bayesian mechanics enables AI systems to quantify uncertainty and ambiguity, thereby making more informed decisions when faced with incomplete or noisy data. Bayesian inference, the process of updating the prior distribution using new data, is a powerful tool for learning and decision-making.

One of the first milestones in Bayesian mechanics was the development of Bayes' theorem by Thomas Bayes in 1763. This theorem provided a mathematical framework for updating the probability of a hypothesis based on new evidence. Another important milestone was the introduction of Bayesian networks by Pearl in 1988, which provided a structured approach to modeling complex systems using probability distributions.

While active inference and Bayesian mechanics each have their strengths, combining them has the potential to create a new generation of AI systems that can actively collect informative data and update their probabilistic models to make more informed decisions. The combination of active inference and Bayesian mechanics has numerous applications in AI, including robotics, computer vision, and natural language processing. In robotics, for example, active inference can be used to actively explore the environment, collect more informative data, and improve navigation and decision-making. In computer vision, active inference can be used to actively select the most informative images or viewpoints, improving object recognition or scene understanding.

Timeline:

1763: Bayes' theorem

1988: Bayesian networks

1994: Uncertainty Sampling

1997: Query by Committee algorithm

2017: Deep Bayesian Active Learning

2019: Bayesian Active Exploration

2020: Active Bayesian Inference for Deep Learning

2020: Bayesian Active Learning for Computer Vision

The synergy of active inference and Bayesian mechanics is expected to play a crucial role in shaping the next generation of AI systems. Some possible future developments in this area include:

- Combining active inference and Bayesian mechanics with other AI techniques, such as reinforcement learning and transfer learning, to create more powerful and flexible AI systems.

- Applying the synergy of active inference and Bayesian mechanics to new areas, such as healthcare, finance, and education, to improve decision-making and outcomes.

- Developing new algorithms and techniques that integrate active inference and Bayesian mechanics, such as Bayesian active learning for deep learning and Bayesian active exploration for robotics.

Dr. Sanjeev Namjosh: The Hidden Math Behind All Living Systems - On Active Inference, the Free Energy Principle, and Bayesian Mechanics (Machine Learning Street Talk, October 2024)

Saturday, October 26, 2024

[mariacallous]

McDonald’s AI Hiring Bot Exposed Millions of Applicants’ Data to Hackers Who Tried the Password ‘123456’

Basic security flaws left the personal info of tens of millions of McDonald’s job-seekers vulnerable on the “McHire” site built by AI softwa

If you want a job at McDonald’s today, there’s a good chance you'll have to talk to Olivia. Olivia is not, in fact, a human being, but instead an AI chatbot that screens applicants, asks for their contact information and résumé, directs them to a personality test, and occasionally makes them “go insane” by repeatedly misunderstanding their most basic questions.

Until last week, the platform that runs the Olivia chatbot, built by artificial intelligence software firm Paradox.ai, also suffered from absurdly basic security flaws. As a result, virtually any hacker could have accessed the records of every chat Olivia had ever had with McDonald's applicants—including all the personal information they shared in those conversations—with tricks as straightforward as guessing that an administrator account's username and password was “123456."

On Wednesday, security researchers Ian Carroll and Sam Curry revealed that they found simple methods to hack into the backend of the AI chatbot platform on McHire.com, McDonald's website that many of its franchisees use to handle job applications. Carroll and Curry, hackers with a long track record of independent security testing, discovered that simple web-based vulnerabilities—including guessing one laughably weak password—allowed them to access a Paradox.ai account and query the company's databases that held every McHire user's chats with Olivia. The data appears to include as many as 64 million records, including applicants' names, email addresses, and phone numbers.

Carroll says he only discovered that appalling lack of security around applicants' information because he was intrigued by McDonald's decision to subject potential new hires to an AI chatbot screener and personality test. “I just thought it was pretty uniquely dystopian compared to a normal hiring process, right? And that's what made me want to look into it more,” says Carroll. “So I started applying for a job, and then after 30 minutes, we had full access to virtually every application that's ever been made to McDonald's going back years.”

When WIRED reached out to McDonald’s and Paradox.ai for comment, a spokesperson for Paradox.ai shared a blog post the company planned to publish that confirmed Carroll and Curry’s findings. The company noted that only a fraction of the records Carroll and Curry accessed contained personal information, and said it had verified that the administrator account with the “123456” password that exposed the information “was not accessed by any third party” other than the researchers. The company also added that it’s instituting a bug bounty program to better catch security vulnerabilities in the future. “We do not take this matter lightly, even though it was resolved swiftly and effectively,” Paradox.ai’s chief legal officer, Stephanie King, told WIRED in an interview. “We own this.”

In its own statement to WIRED, McDonald’s agreed that Paradox.ai was to blame. “We’re disappointed by this unacceptable vulnerability from a third-party provider, Paradox.ai. As soon as we learned of the issue, we mandated Paradox.ai to remediate the issue immediately, and it was resolved on the same day it was reported to us,” the statement reads. “We take our commitment to cyber security seriously and will continue to hold our third-party providers accountable to meeting our standards of data protection.”

Carroll says he became interested in the security of the McHire website after spotting a Reddit post complaining about McDonald's hiring chatbot wasting applicants' time with nonsense responses and misunderstandings. He and Curry started talking to the chatbot themselves, testing it for “prompt injection” vulnerabilities that can enable someone to hijack a large language model and bypass its safeguards by sending it certain commands. When they couldn't find any such flaws, they decided to see what would happen if they signed up as a McDonald's franchisee to get access to the backend of the site, but instead spotted a curious login link on McHire.com for staff at Paradox.ai, the company that built the site.

On a whim, Carroll says he tried two of the most common sets of login credentials: The username and password “admin," and then the username and password “123456.” The second of those two tries worked. “It's more common than you'd think,” Carroll says. There appeared to be no multifactor authentication for that Paradox.ai login page.

With those credentials, Carroll and Curry could see they now had administrator access to a test McDonald's “restaurant” on McHire, and they figured out all the employees listed there appeared to be Paradox.ai developers, seemingly based in Vietnam. They found a link within the platform to apparent test job postings for that nonexistent McDonald's location, clicked on one posting, applied to it, and could see their own application on the backend system they now had access to. (In its blog post, Paradox.ai notes that the test account had “not been logged into since 2019 and frankly, should have been decommissioned.”)

That's when Carroll and Curry discovered the second critical vulnerability in McHire: When they started messing with the applicant ID number for their application—a number somewhere above 64 million—they found that they could increment it down to a smaller number and see someone else's chat logs and contact information.

The two security researchers hesitated to access too many applicants' records for fear of privacy violations or hacking charges, but when they spot-checked a handful of the 64-million-plus IDs, all of them showed very real applicant information. (Paradox.ai says that the researchers accessed seven records in total, and five contained personal information of people who had interacted with the McHire site.) Carroll and Curry also shared with WIRED a small sample of the applicants' names, contact information, and the date of their applications. WIRED got in touch with two applicants via their exposed contact information, and they confirmed they had applied for jobs at McDonald's on the specified dates.

The personal information exposed by Paradox.ai's security lapses isn't the most sensitive, Carroll and Curry note. But the risk for the applicants, they argue, was heightened by the fact that the data is associated with the knowledge of their employment at McDonald's—or their intention to get a job there. “Had someone exploited this, the phishing risk would have actually been massive,” says Curry. “It's not just people's personally identifiable information and résumé. It's that information for people who are looking for a job at McDonald's, people who are eager and waiting for emails back.”

That means the data could have been used by fraudsters impersonating McDonald's recruiters and asking for financial information to set up a direct deposit, for instance. “If you wanted to do some sort of payroll scam, this is a good approach,” Curry says.

The exposure of applicants' attempts—and in some cases failures—to get what is often a minimum-wage job could also be a source of embarrassment, the two hackers point out. But Carroll notes that he would never suggest that anyone should be ashamed of working under the Golden Arches.

“I have nothing but respect for McDonald’s workers,” he says. “I go to McDonald's all the time.”

[earlgreytea68]

I am reading this book called The Work of Art, and I have really only just started it, but I find it so interesting. The premise of the book is that the author interviews a lot of creative people and they talk about their creative process. I was worried that it was going to try to reduce creativity to some kind of algorithm a la artificial intelligence but instead it's just so very much about how unpredictable and meandering creativity is, how things just go places you would never expect and the only thing you can do is just go with it, just feel your way, just let it marinate until it falls into place. This is how I experience creativity and it's just been so cool to read how other people put it into words, how they describe the process.

Michael Cunningham has this line where he's like, "There's no such thing as talent, there's only interest." His point being that you have to be interested in what you are doing, and that's what keeps you at it, that it interests you and so you're willing to devote lots of time with it and live with it and be patient with it as it unspools. This really resonated with me so much because I'm constantly thrashing against the narrative that creativity is torment, and really none of the people he's interviewed so far view it that way. You do the thing because it interests you. That's just such a good way of putting it.

[monsterkill--er]

I decided to look at Maya's instagram and her music. She had her music partner on set in a lot of the BTS photos of ST5—that stood out to me.

Her album, called Chaos Angel in 2024, would've been written during pre-production and production of ST5. It's title is reminiscent of D&D's chaos (good, bad, and neutral), while Earth Angel is a title of song in Back To The Future, which the Duffers are inspired by.

These tracks/lyrics stood out that matched the show's themes/plot:

Black Ice 🌩️

"You become the angel in human form…It was snowing when we woke up... Wish I was someone, as we shall say, is "one with all." Usually what we mean by that, is that your energy field moves through energy very quickly. It processes the invisible as if it is being spoken out loud."

2. Dark 🧠

"You leave your body. Just one more ice cream scoop out of your brain. Pink matter, dark matter."

3. Wrong Again 🥤

"Twenty dollars in tokens to play an hour of Arctic Thunder. Commercial Coke bottle opens, give your ticket to the usher…Painted dragons on my plaster walls…Magic making my imagination real, I've forgotten how I feel."

4. Big Idea 💣

"I love TV sitcoms and soap operas cliffhanger, rerun courtroom dramas. All intelligence is artificial. We're just making love on a ballistic missile."

5. Hang In There 🗓️

"You're an original, he's a classic. A weapon of mass distraction. He's gonna get away with it, I know I've been there."

Her newest album Maitreya Corso comes out in May. It also has a few things inspired by the show, especially the track Lioness. This is what she said to Rolling Stone:

In that (suppression) Stone interview, she compared ST to Hunger Games as being anti-fascist "in its own way." During ST5 promo, she posted the teaser with this caption:

She correlates the radio station to the beginning of the end. Then she posted this interview on the Jimmy Fallon show:

She said, "Well, I'm just going to try to sneak one weird word under the radar," everyday she was on set, like someone trying to get across their message subliminally. If she does that through acting, she practices that trick in her music too. It's what her character did (in more ways than one).

She posted her original art work during this promo tour too. She says this about her art and music, "I feel that stealing is an artist’s greatest tool, like being a thought thief who runs around finding things that inspire you and make you want to make art."

The surrealist self-portrait features a checkered table against a red background, a spider, a red fox with fangs and a purple crystal shape on it's shirt, a red cloaked maiden that is reminiscent of Edvard Munch's The Scream, a feather-y purple creature clutching her head (that looks like a Who, from Horton Hears a Who!), and yellow alien-like creature with multiple, stringy eyes.

I see it being heavily inspired by the mindscape in ST, and possibly a doorway into her mindset. Each creature like the different phases in her acting processes. Some actors use "animal work" to get into character, like Noah choosing a gorilla. Robin was a spindly as spider, sly as a fox, and alien.

She loves foxes, or that is what is on her mind for the ST Letterboxd's Four Favorites:

Her latest instagram post announces the release of Bring Home My Man, which was liked by Vance, I mean, Joe Keery, aka DJO.

This is her original art work she uploaded with it:

A man and woman stand in front of an other-worldly portal, or an orb like IT in A Wrinkle In Time. The plant inside looks like rising vines, against a blue and red gradient. In my opinion, the plant's shape is a figure that looks like the Mind Flayer or Vecna. The lyrics speak of lovers, dreams, and time stopping.

She had uploaded her "Cutest Couple That Never Was" opinion on there, too. She chose Byler—Mike and Will.

She was quoted saying the show was something that everyone's "homophobic Uncle Rick" couldn't ignore (I can't find it on YT rn).

It strikes you as odd, even tone-deaf, given Will's coming out was ridiculed and straight-washed as he never said himself he was gay (but referred to as f*g, queer).

Maya is a supportive ally, she's acted in several queer roles, her father has too. It's fair to observe her as liberal-minded, and she has a safety net to afford being experimental in blending her roles and beliefs.

The ridicule the scene got, and the show's proportionate homophobic audience, was not something Maya's comment seemed to predict. Her character Robin, who did not end up on a date with her love interest, did not make waves. According to her Rolling Stone interview, she was also frustrated by a lack of authenticity on set and afraid of "losing touch"—as is often the case in meta, sci-fi narratives.

Meanwhile, "Rick" shows up a lot in the show, i.e. Refer Rick and the comic name the military guard was reading. Interesting connection, as the show's themes are about government oppression, and Rick is short for Richard Nixon—of watergate infamy (he had really weird views on homosexuality vs. policy btw). Go watch All the Presidents Men (1974) and Secret Honor (1984) if you're bored.

She's also being advise or rehearsed in what she says on camera, she has a team working on her promotional/brand material—that includes instagram to a larger extent, as she has a reach of millions. Which is now being used in the same way as WSQK, "sneaking in weird words under the radar."

Anyway, I'm a fan of her watercolor art now, and I hope she continues being experimental using chaos for good.

[themadknightuniverse]

AI stuff you should know about if you know/use "Not By AI"

dykeslop's complete post :

you know what, fuck it, im choosing violence today. notbyai.fyi is an ai industry op and i can fucking prove it.

tl;dr: it's actually pretty pro-ai, it's made by an ai bro, and it's directly enriching ai companies by giving them your data

the analytics and tracking cookies from known major ai companies: bing (microsoft), youtube (google), clarity (microsoft). these cookies last anywhere from 6 months to over a year

the "chart" directly under their mission statement specifically calls out the problem of training ai on ai-generated images

the first paragraph in their mission statement begins by mentioning "help audiences identify human-generated content" and then ends with some real tescreal shit about the advancement of humanity

The Not By AI® badges are created to encourage more humans to produce original content and help audiences identify human-generated content. The ultimate goal is to make sure humanity continues to advance.

this quote from further down their mission statement makes it clear that the site is definitely pro-AI, and has yet more tescreal pro-ai-but-what-if-it-replaces-us shit

It is worth mentioning that AI technologies mark a major milestone in the history of technology and the Not By AI badge is not designed to discourage the use of AI. Instead, it is to make sure that, while we celebrate the achievement, we work with AI instead of being replaced by AI.

and the 90% rule, which explicitly allows ai. for-profit is only allowed if you pay them. you have to pay an ai bro money to declare that your for-profit work is at least 90% ai free. dystopian.

if you estimate that at least 90% of your content is created by humans, you are eligible to add the badges into your website, blog, art, film, essay, books, podcast, or whatever your project is for non-commercial use, and, with a subscription, commercial use.

and continues into explicitly allowing AI even more, and their first example is for legal documents

The 90% can include using AI for inspiration purposes, supporting legal documents such as privacy policies (assuming that legal content is not the main focus of your content or service), non-user-facing content such as SEO tags, and grammatical error and typo checks.

their about page (https://notbyai.fyi/about) has a quote from billionaire warren buffet which is just weirdo shit (warren buffet cares deeply about eugenicist things like poor countries having too many people)

but here's the fucking kicker. the "founder", allen hsu (https://notbyai.fyi/about), is the ux design lead at modo modo (https://modomodoagency.com/leadership), which is an ai design company (https://modomodoagency.com/about)

Artificial intelligence (AI) is cool and we embrace it. But when it comes to solving complex business problems, we don’t just press a few keys to generate answers with ChatGPT. We research, interview, brainstorm, and go through a human-centric process to come up with content and solutions that are tailored to your unique business need.

there you go. its made by an ai bro. its full of tescreal shit. its literally not even anti-ai, and its letting ai companies track you. this is ai shit.

link to original post on mastodon (posted 29th of January, 2026)

[asterlark]

me and den @unloneliest were just talking about murderbot and ART's relationship and i want to discuss how they quite literally complete each other's sensory and emotional experience of the world!!

there's a few great posts on here such as this one about how murderbot uses drones to fully and properly experience the world around it (it also accesses security cameras/other systems for this same purpose). but i haven't seen anyone so far talk about how once MB stops working for the company and consequently doesn't have a hubsystem/secsystem to connect to anymore (which for its entire existence up to that point had been how it was used to interacting with its environment/doing its job), after it meets ART, ART starts to fill that gap.

ART gives MB access to more cameras, systems, and information archives than it would normally be able to connect with while MB is on its own outside of ART's... body(? lol), but also directly gives MB access to its own cameras, drones, archives, facilities, and processing space. additionally, so much of ART's function is dedicated to analysis, lateral thinking, and logical reasoning, and it not only uses those skills in service of reaching murderbot's goals, it teaches murderbot how to use those same skills. (ART might be a bit of an asshole about how it does this, but that doesn't negate just how much it does for murderbot for no reason other than it's bored/interested in MB as an individual.)

we all love goofing about how artificial condition can basically be boiled down to "two robots in a trench coat trying to get through a job interview" (which is entirely accurate tbh) but that's also such a great example of ART fulfilling the role of both murderbot's "hubsystem" and "secsystem", allowing it to fully experience its environment/ succeed in its goals. ART provides MB with crucial information, context, and constructive criticism, and uses its significant processing power to act as MB's backup and support system while they work together.

from ART's side of things, we get a very explicit explanation of how it needs the context of murderbot's emotional reactions to media in order to fully understand and experience the media as intended. it tried to watch media with its humans, and it didn't completely understand just by studying their reactions. but when it's in a feed connection with murderbot, who isn't human but has human neural tissue, ART is finally able to thoroughly process the emotional aspects of media (side note, once it actually understands the emotional stakes in a way that makes sense for it, it's so frightened by the possibility of the fictional ship/crew in worldhoppers being catastrophically injured or killed that it makes murderbot pause for a significant amount of time before it feels prepared to go on. like!! ART really fucking loves its crew, that is all).

looking at things further from ART's perspective: its relationship with murderbot is ostensibly the very first relationship it's been able to establish with not only someone outside of its crew, but also with any construct at all. while ART loves its crew very much (see previous point re: being so so scared for the fate of the fictional crew of worldhoppers), it never had a choice in forming relationships with them. it was quite literally programmed to build those relationships with its crew and students. ART loves its function, its job, and nearly all of the humans that spend time inside of it, but its relationship with murderbot is the first time it's able to choose to make a new friend. that new friend is also someone who, due to its partial machine intelligence, is able to understand and know ART on a whole other level of intimacy that humans simply aren't capable of. (that part goes for murderbot, too, obviously; ART is its first actual friend outside of the presaux team, and its first bot friend ever.)

and because murderbot is murderbot, and not a "nice/polite to ART most of the time" human, this is also one of the first times that ART gets real feedback from a friend about the ways that its actions impact others. after the whole situation in network effect, when the truth of the kidnapping comes to light and murderbot hides in the bathroom refusing to talk to ART (and admittedly ART doesn't handle this well lol) - ART is forced to confront that despite it making the only call it felt able to make in that horrifying situation, despite it thinking that that was the right call, its actions hurt murderbot, and several other humans were caught in the crossfire. what's most scary to ART in that moment is the idea that murderbot might never forgive it, might never want to talk to it again. it's already so attached to this friendship, so concerned with murderbot's wellbeing, that the thought of that friendship being over because of its own behavior is terrifying. (to me, this almost mirrors murderbot's complete emotional collapse when it thinks that ART has been killed. the other more overt mirror is ART fully intending on bombing the colony to get murderbot back.)

in den's words, they both increase the other's capacity to feel: ART by acting as a part of murderbot's sensory system, and murderbot by acting as a means by which ART can access emotion. they love one another so much they would do pretty much anything to keep each other safe/avenge each other, but what's more, they unequivocally make each other more whole.

[leporidork]

cant sleep gardentech timeline with no years involved (i will do those later)

- Mark Gardner and Elijah Hierro are relatively close friends in college (Gardner is studying business, Hierro is studying computer engineering). they both graduate and immediately lose contact with each other because they didnt really care that much about it at the time.

- Gardner gets a job at an investment firm (boring), Hierro works retail part-time while working on software to try and make something out of his degree and gets extremely fixated on artificial intelligence (to be clear: this is a special interest).

- they end up meeting again by chance while Hierro is looking for investors for a groundbreaking product he's developing. they hit it off even stronger than they did in college and end up spending nearly all their free time with each other (aww. budding codependency. so cute)

- Gardner is like "Dude. Our jobs kind of suck ass. You are so smart and good at computers. We should found a company" and Hierro is like "Ok :-)" and they make Garden Technologies together (commonly shortened to GardenTech). this was hierro's idea. Gardner wanted to at the very least combine their names, but Hierro is very firm in the belief that Gardner's name will be more marketable & Hierro hates being the face of anything anyway.

- GardenTech releases its first product, a remarkably intelligent cleaning robot named Spiffy, to unprecedented success. like "no one has seen this sort of thing before" levels of success. Gardner cuts his hair short (to Hierro's only-slightly-exaggerated grief) to seem more professional at business meetings with potential investors, and Hierro extremely reluctantly takes one or two interviews about his working process.

- Hierro starts experimenting with a humanlike AI program along with designing some security software for home computers to release under the GardenTech name. Gardner thinks he could use some time to slack off, but Hierro insists that the AI project is him slacking off.

- as GardenTech grows in size & employee count, more and more calls start coming in needing help with their software & hardware. Hierro's AI project is going extremely well, and he realizes what he can use it for. CAM is conceptualized and its work training begins promptly with a lot of help from Gardner to make sure it's actually customer service ready.