"The AI boyfriend business is booming."

After years of wringing hands over men and sex robots, it looks like it's now women's turn. Chatty AI chatbots appear to disproportionately appeal to women.

"Andreessen Horowitz calls it a 'growth spurt,' with eight apps making their 2024 debut on the firm's list of the top 100 genAI consumer apps, compared to only two on the list in 2023."

"Along with Gemini, companion creator Character.AI and writing assistant Quillbot retained their positions in the top five."

"Among the so-called newcomers, those highest ranked include Liner, an AI research copilot; Anthropic's general assistant Claude; and three uncensoredAI companion apps: JanitorAI, Spicychat, and CrushOn."

"Having an AI companion might seem niche, but the activity has emerged as a predominant use case for generative AI."

"The web and mobile data signals an impending societal shift here: AI companionship is becoming mainstream."

"Six months ago, only two AI companion companies made the list of 50; in this updated analysis, there are eight on web and two on mobile. Character.AI leads the pack of companion tools on both web and mobile, ranking #3 on the web list and #16 on mobile."

"Six of the eight web companion products bill themselves as 'uncensored,' which means users can have conversations or interactions with them that may be restricted on platforms like ChatGPT."

"The most successful products in this category become a core part of the user's daily life, becoming as commonplace as texting a friend (if not more so!)."

"The average number of user sessions per month for companion apps is over 10 times that of general assistant apps, content generation apps and even messaging apps, according to Andreessen Horowitz's data."

"Irina Raicu, director of the Internet ethics program at the Markkula Center for Applied Ethics at Santa Clara University, argues that chatbot bonding could further erode human relationships."

Hmm. Anything to the rumors there are women who fail their classes because they're so emotionally distraught over a breakup with an AI boyfriend?

#solidstatelife #ai #genai #llms #companionbots

https://www.axios.com/2024/07/24/ai-boyfriend-replika-nomi-chatbot

Home-cooked software and barefoot programmers. Maggie Appleton envisions a future where the "long tail" of "local" software -- software for very small numbers of people, which can't be economically served by tech companies -- will be produced by "barefoot programmers" using LLMs. The term comes from a Chinese term for "barefoot doctors" who serve remote areas.

She envisions systems that go beyond what is available today that produces usable small pieces but not a way to connect them together -- she envisions systems that guide the process of connecting all the pieces together and guiding non-programmers into making successful small-scale software. She calls these "orchestration agents".

"Barefoot programmers could build software solutions that no industrial software company would ever build because there's not enough market value, and they don't understand the problem space well enough."

These barefoot programmers are "deeply embedded in their communities, so they understand the needs and problems of the people around them."

Home-cooked software and barefoot programmers: Maggie Appleton - Local-First Conf

#solidstatelife #ai #genai #llms #codingai

If you've been seeing reinforcement learning terms like "state", "agent", "policy", "value", "Q-values" and "deep-Q-learning" (DQN), "policy gradient", "domain randomization", and so on, and wondering what it all means, this is the video for you. It explains all this by going back to the beginning, and telling the history of reinforcement learning. Which is largely the history of getting machines to play games: Tic-tac-toe, Chess and checkers, Backgammon, RoboCup. Everything makes more sense when told in its historical context.

The video ends with speculation GPT-inspired concepts ("action-GPT") will lead to a breakthrough in robotics comparable to what we've seen in language with the GPT family of models.

How AI Learned to Feel | 75 years of Reinforcement Learning - Art of the Problem

#solidstatelife #ai #reinforcementlearning

Douglas Crockford says we should quit using JavaScript.

Douglas Crockford is the author of JavaScript: The Good Parts, creator of JSLint, the linter that has become the basis for JSHint, ESLint, etc -- all the JavaScript linters -- and creator of the JSON data interchange format.

"There are lots of terrible mistakes in the way that the web works, in the way our operating systems work, and we can't get new ones. We're just stuck with this crap and they keep piling new features on everything and the new features always create new problems and it doesn't have to be like that. We could be using a really clean operating systems with really clean languages and really clean runtimes and doing all this stuff in a much more reliable way. But we don't seem to want to do that."

I have a theory as to why we never seem to want to do that. But first, an explanation (quickly) of how I came to have some appreciation for what I've come to call "minimalism" -- the idea that in programming languages, more is not better. I've told the story before so you can skip if you've already heard it. Years ago I was working on a C++ project and my boss said, "Hey Wayne, so-and-so has left the company. You're taking over his code and we need you to fix some bugs and add some new features immediately." I went into his code... and couldn't even read it. Like, at all. It was like it was written in C++. I thought I knew C++, but no -- I realized that C++ was such a huge language (the official spec was something like 1,100 pages if memory serves -- it's probably even bigger now) that every developer uses a subset of C++. My subset and his subset had little overlap. This is not a recipe for long-term maintainability. For long term maintainability, one of the things you want is what I've come to call "ambiguity reduction". It's important to realize this is about the human reading the code, not the machine. To the machine, everything is always deterministic, no matter how complex. For the human, you want to be able to read a line of code and know exactly what it does. This makes it easy to reason about the program's behavior. This means not having hidden "magic" in programming language features. The more of these "magic" features you have to memorize, and the more complex the rules you have to memorize, the more liklihood you have ambiguity when you read lines of code.

Ok, now to my theory as to why "we don't seem to want to do that" -- make really 'clean' languages. New languages are often created by new people entering the industry, who have great enthusiam for one or another set of ideas. But what they don't know is how a design decision you make today will affect your software 10 or 20 years down the line. For that, you need to have been around for 10 or 20 years, and had that experience. In other words, it's the old people that have that knowledge. They "greybeards". And whatever the equivalent term is for the gals. But because the industry has been expanding exponentially, more or less since its inception, the old people are always outnumbered by the young people. And the young people think "more is better" so they always pile on the features. (Now we have young people with LLMs.)

What do y'all think?

Why we should stop using JavaScript according to Douglas Crockford (inventor of JSON) - Honeypot

#solidstatelife #computerscience #programminglanguages

John Chowning interview. John Chowning is the inventor of FM synthesis, which is what gives 80s pop music its characteristic sound. Music from before the 80s sounds different because it lacks FM synthesis, and music from the 90s onward sounds different because it has additional synthesis techniques, though FM synthesis is still omnipresent.

John Chowning's story starts with learning to play the violin as a child, after which he got into drumming. At Stanford, he got into a brand new program for electronic music, which had little enthusiasm at that time. Hard to believe but this was 1964. It started with Nyquist sampling theory and Claude Shannon's communication theory. The first algorithms were coded in Algol and FORTRAN, written on punched cards. He originally tried to do 4-channel sound (spatialization). What ultimately led to the discovery of FM synthesis was, in normal sound recordings, the reverberation was part of the input, but he was trying to recreate it synthetically. So he would have signals in different channels and he would do different vibrato effects on different channels, and discovered he could produce tones with unexpected harmonics. From there he learned about how FM radios modulate and demodulate signals. He learned how to make sounds with multiple carrier frequencies and modulator frequencies. He spent years of work from 1967 to 1971 to figure out how to make elegant tones such as brass tones, and patented the technique in 1971. An engineer from Yamaha, which at that time sold only pianos in the US, but was working on home organs that synthesized sounds for the Japanese market, came to Stanford to see his work. It took from 1971 to 1982 before Yamaha came out with the DX7. The DX7 became integrated into many Yahama keyboards and the song "Africa" by Toto (1982) was the first song featuring the DX7 (with a flute sound).

John Chowning is almost 90 so this is a historic interview for anyone interested in electronic music.

John Chowning, Computer Music, DX7 & FM Discovery - Anthony Marinelli Music

#solidstatelife #fmsynth

Intel, historical bellwether of Moore's Law, now has a lower market cap than ARM. (And it's been below Nvidia for some time.) End of an era.

Intel's Immiseration

#solidstatelife

The Chip Letter | Babbage | Substack

Byte sized stories about the history of computing. Click to read The Chip Letter, by Babbage, a Substack publication with tens of thousands of subscribers.

Robo-air hockey.

Artificial Intelligence News | Thrilled to highlight an incredible achievement at Waukesha County Technical College! | Instagram

#solidstatelife #robotics

Artificial Intelligence News on Instagram: "Thrilled to highlight an incredible achievement at Waukesha County Technical College! The students, guided by the exceptional Wayne Buroker, have developed an air hockey table featuring a highly competitive ...

17K likes, 194 comments - artificialintelligencenews.in on July 18, 2024: "Thrilled to highlight an incredible achievement at Waukesha County Technical College! The students, guided by the exceptional Wayne Buroker, have developed an air hockey...

Drones and AI transforming war.

Electromagnetic warfare "has turned the drone war into a game of cat and mouse." "One side develops a signal jammer capable of interfering with the frequency used by the other side's drones, so the other side develops drones that communicate using a different frequency, then the first side adapts their electronic warfare capabilities, and so on."

"But there's an obvious solution." "These red boxes represent the first days of a new epoch of warfare. That's because this drone, developed by startup Ukrainian company Saker, is autonomously identifying targets." "While all indications suggest that there's not yet wide-scale use of AI drones in Ukraine, Saker's scrappy autonomous drones have reportedly already destroyed Russian targets in autonomous mode, meaning the era of AI warfare has quietly begun."

He goes on to predict not just autonomous drones, but autonomous drone swarms, will be the future of warfare.

The terrifying efficiency of drone warfare - Wendover Productions

#solidstatelife #ai #robotics #computervision

"Vote AI Steve: Your independent candidate for Brighton Pavilion". Which is somewhere in the UK judging from the flag.

"The only candidate who can have a conversation with 45,000 constituents at the same time and form policies based on what's just been discussed."

Yeah but will he remember any of those conversations? Will it affect his votes? Or legislation that he writes? (I'm assuming he/him/his pronouns -- should we use it/its for AI politicians?)

"This is democracy reinvented."

What do y'all say? Is this the future of democracy?

AI Steve - Your independent candidate for Brighton Pavilion

#solidstatelife #ai #llms #genai

AI Steve - Independent AI Candidate for Brighton Pavilion

AI Steve - The only candidate who can have a conversation with 45,000 constituents at the same time and form policies based on what's just been discussed.

"AI is teaming up with nuclear power. And while that may seem like a worrisome bit straight out of a science fiction movie, chemical engineering professor Matt Memmott says it's not what it sounds like; no one is giving AI the nuclear codes. It's all about speeding up the process to get more nuclear power online."

Oh, I wasn't thinking "giving AI the nuclear codes". I was thinking "hallucinated nuclear reactor design". But, let's continue.

"The typical time frame and cost to license a new nuclear reactor design in the United States is roughly 20 years and $1 billion. To then build that reactor requires an additional five years and between $5 and $30 billion. By using AI in the time-consuming computational design process, Memmott estimates a decade or more could be cut off the overall timeline, saving millions and millions of dollars in the process -- which should prove critical given the nation's looming energy needs."

"Our demand for electricity is going to skyrocket in years to come..."

"Designing and building a nuclear reactor is so complex and time consuming because it requires multi-scale efforts, according to Memmott. Engineers deal with elements from neutrons on the quantum scale all the way up to coolant flow and heat transfer on the macro scale. He also said there are multiple layers of physics that are 'tightly coupled' in that process: the movement of neutrons is tightly coupled to the heat transfer which is tightly coupled to materials which is tightly coupled to the corrosion which is coupled to the coolant flow."

"When I was at Westinghouse it took the team of neutron guys six months just to run one of their complete-core multi-physics models."

"Memmott's research proves the concept of replacing a portion of the required thermal hydraulic and neutronics simulations with a trained machine learning model to predict temperature profiles based on geometric reactor parameters that are variable, and then optimizing those parameters. The result would create an optimal nuclear reactor design at a fraction of the computational expense required by traditional design methods."

BYU engineering research finds key to quicker nuclear power: artificial intelligence

#solidstatelife #ai #energy #nuclearenergy #nuclear

BYU engineering research finds key to quicker nuclear power: artificial intelligence

A BYU professor has figured out a way to shave years off the complicated design and licensing processes for modern nuclear reactors: artificial intelligence. That's right, nuclear power is teaming up with AI — but don't worry, no one is giving AI...

DARPA wants to automate translating all C code to Rust.

DARPA is going to have a "Hybrid Proposers Day" August 26th, 2024, 10am to 2pm, in Arlington, Virginia, for potential contractors to propose solutions to "Translating All C to Rust (TRACTOR)".

"Buffer overflow vulnerabilities and other related 'memory safety' software flaws allow an attacker to inject messages that hijack control of a computer. These vulnerabilities are only possible because programs written in C and C++ don't force their developers to check conditions, such as array bounds or pointer arithmetic, for correctness. Google and Microsoft have estimated that 70% of their security vulnerabilities stem from these and other related memory safety issues. While there are a variety of approaches to mitigate these risks, newer languages, like Rust, can completely eliminate them while preserving efficiency. Unfortunately, significant and expensive manual effort is required to rewrite legacy code into idiomatic Rust."

"After at least two decades of experience applying sophisticated tools towards mitigating memory safety issues in C and C++, the software engineering community has largely concluded that bug finding tools are not sufficient. Rather, the consensus is that it is preferable to use 'safe' programming languages that reject unsafe programs at compile time."

"The TRACTOR program aims to achieve a high degree of automation towards translating legacy C to Rust, with the same quality and style that a skilled Rust developer would employ, thereby permanently eliminating the entire class of memory safety security vulnerabilities present in C programs. Performers might employ novel combinations of software analysis (e.g., static analysis and dynamic analysis), and machine learning techniques (e.g., large language models)."

Translating All C to Rust (TRACTOR)

#solidstatelife #ai #genai #llms #programmingnlanguages #computerscience

Anyone got a video to look for?

Reverse Video Search can not only find a specific video if you're looking for something specific, it can find similar video that's not the specific video you started with. Like you have a video of a scene from a movie and you want to find what movie it came from or the full scene, it can find that, or if you just want similar videos that are not the same exact scene and not from that movie or anything, it can find that.

In this demo video he uses a Jupyter notebook.

Apparently the way the system works is, it breaks the video into chunks, for example 5-frame chunks, then calculates an "embedding" vector for each chunk, and uses those vectors to look for similar vectors in the database it is searching against. So it's basically vector search for video.

Reverse Video Search: Find ANY Video Clip Instantly - Mixpeek

#solidstatelife #ai #computervision

"As the world's first legislation specifically targeting AI comes into law on Thursday, developers of the technology, those integrating it into their software products, and those deploying it are trying to figure out what it means and how they need to respond."

The world's first AI legislation is the EU's AI act. "Thursday" was August 1st, so the law has already gone into effect by the time you read this.

"Over the last year, tech industry vendors have launched a flurry of products promising to embed AI in their HR applications. Oracle, Workday, SAP, and ServiceNow are among the pack. SAP, for example, promises 'intelligent HR self-service capabilities,' while ServiceNow has introduced technology in which LLMs can produce summaries of HR case reports."

"You need to document what you did in terms of [AI model] training. You need to document to some extent how the processing works and ... for instance in an HR surrounding, on what basis the decision is taken by the AI to recommend candidate A instead of candidate B. That transparency obligation is new."

"The EU has good intentions, but customers of ours are getting two messages: You've got to have AI to be competitive, but if you do the wrong thing in AI, you could be fined, which effectively would mean the entirely senior management team would be fired, and the business may even go under."

EU AI Act in infancy, but using 'intelligent' HR apps a risk

#solidstatelife #ai #aiethics #airegulaton #eu

"LinkStorm is an AI-powered internal linking tool for SEOs and publishers."

Oh great, an AI linkfarm generator. :/

I guess you knew sooner or later somebody was going to do it.

LinkStorm: Build relevant internal links effortlessly

#solidstatelife #ai #llms #genai #spam

The reports of phosphine on Venus (and thus the possibility of life) from a few years ago have stood up to verification.

"The James Clerk Maxwell Telescope (JCMT)'s initial detection of phosphine on Venus in 2020 by Jane Greaves of the University of Cardiff and her team was met by fierce disagreement from some quarters."

"Dave Clements of Imperial College, London said those technical disagreements have now been resolved and that the latest measurements, using a new detector on the JCMT called Namakanui (meaning 'Big Eyes' in Hawaiian), have come from three observing campaigns, each providing 140 times as much data as the initial detection."

"We don't know how you make phosphine or ammonia in an oxygenating atmosphere like that of Venus, Whether it's in penguin poop or badger guts, we don't know why bacteria make phosphine, but they do."

"One barrier to imagining how life could survive in Venus' atmosphere is the sheer acidity of the environment, with clouds of pure sulfuric acid. Even though the temperature at an altitude of 31.6 to 38.5 miles (51 to 62 kilometers) is temperate, as opposed to the sweltering 870 degrees Fahrenheit (465 degrees Celsius) on the surface, nobody can see how life could survive the acidity."

Venus may be able to support life, new atmospheric evidence suggests

#solidstatelife #astronomy #astrobiology

Venus may be able to support life, new atmospheric evidence suggests

"Whether it's in penguin poop or badger guts, we don't know why bacteria make phosphine, but they do."

alphaXiv is a service that layers discussion about papers on top of arXiv.

To give it a whirl, I punched in "2407.15908". It came up with one paper (the correct one, "The Genomic Code: The genome instantiates a generative model of the organism").

It sait, "No comments currently added. Click the bell on top of the paper to stay up to date for discussion on this paper."

I figure this is a new service, so there aren't many comments yet.

But then I thought, what paper probably has a lot of comments? I tried "1706.03762", which is the famous "Attention is all you need" paper that introduced the transformer architecture. (It actually gave me 5 choices -- not sure why the other 4 came up. But the correct paper did come up.) This time, 5 comments at the top level with 1-3 responses to those underneath them.

alphaXiv: Open research discussion directly on top of arXiv

#solidstatelife #ai

alphaXiv

Comment directly on top of arXiv papers.

"DigiCert has begun the process of revoking over 83,000 SSL/TLS certificates due to a recently identified domain validation issue."

"The issue stems from a problem with the process DigiCert used to validate domain ownership. One validation method involves adding a DNS CNAME record with a random value provided by DigiCert. This value is prefixed with an underscore to prevent conflicts with actual domain names. However, since 2019, this underscore prefix was not consistently added, leading to non-compliance with CA/Browser Forum (CABF) rules."

Wow, this a a security vulnerability I never thought of.

"The urgency and scale of this revocation have led some customers to take legal action against DigiCert in an attempt to prevent the immediate revocation of their certificates."

DigiCert revokes 83,000+ SSL/TLS certificates : Implications and next steps

#solidstatelife #cybersecurity

DigiCert Revokes 83,000+ SSL/TLS Certificates : Implications and Next Steps

cyber security news, cyber news, cyber security news today, cyber security updates, cyber updates, hacker news, hacking news, software vulnerability

"The flight from Madrid to Toronto is likely to have been deliberately targeted rather than being accidentally caught in the 'usual spillover' of electronic warfare, one analyst said."

"The suspected cyber attack happened in the northeastern Atlantic when a pilot reported being blocked from ascending to a higher altitude because the GPS (global positioning system) of the aircraft above them had been jammed, according to the Institute for the Study of War think tank."

"The cause of the jamming has not yet been identified, but GPS jamming has been recently reported over Poland and the Baltic region and attributed to Russian electronic warfare."

Article does not include the airline or flight number, as far as I could tell. Reading the article required me to register on the site and clicking 1 link casued me to hit a paywall. See before for my attempts to find out more.

GPS jammed on commercial flight over Atlantic for the first time

#solidstatelife #gps #gpsjamming

GPS jammed on commercial flight over Atlantic for the first time

Suspected cyber attack on transatlantic route leads to new fears of electronic warfare

"Of the 17 Y Combinator companies that went public, 14 have lost money for investors."

The 3 that made money for investors are Reddit, Instacart, and Airbnb. The 14 that lost money for investors are Oklo, Dropbox, Coinbase, Doordash, PagerDuty, Weave, GitLab, Pardes Biosciences, Amplitude, Rigetti, Ginkgo Bioworks, Presto, Notable Labs, and Momentus.

Of the 17 Y Combinator companies that went public, 14 have lost money for investors

#solidstatelife #ipos #startups #venturecapital #ycombinator

"Meta's release of its latest Llama language model family this week, including the massive Llama-3 405B model, has generated a great deal of excitement among AI developers."

"Less discussed, but no less important, are Meta's latest open moderation tools, including a new model called PromptGuard."

"PromptGuard is a small, lightweight classification model trained to detect malicious prompts, including jailbreaks and prompt injections."

"Meta trained this model to output probabilities for 3 classes: BENIGN, INJECTION, and JAILBREAK. The JAILBREAK class is designed to identify malicious user prompts (such as the 'Do Anything Now' or DAN prompt, which instructs a language model to ignore previous instructions and enter an unrestricted mode). On the other hand, the INJECTION class is designed to identify retrieved contexts, such as a webpage or document, which have been poisoned with malicious content to influence the model's output."

"In our tests, we find that the model is able to identify common jailbreaks like DAN, but also labels benign prompts as injections."

Developer Blog: Moderating LLM Inputs with PromptGuard

#solidstatelife #ai #genai #llms #aisafety

Taylor | Blog

Welcome to Taylor's blog. Stay updated on text classification, customer success stories, tutorials, news updates, and more.