Artificial intelligence is rapidly changing many aspects of how we work and live. (How many stories did you read last week about self-driving cars and job-stealing robots? Perhaps your holiday shopping involved some AI algorithms, as well.) But despite the constant flow of news, many misconceptions about AI remain.
AI doesn’t think in our sense of the word at all, Scriffignano explains. “In many ways, it’s not really intelligence. It’s regressive.”
IT leaders should make deliberate choices about what AI can and can’t do on its own. “You have to pay attention to giving AI autonomy intentionally and not by accident,”
6,044 Comments
Tomi Engdahl says:
Rabbitt.ai Launches ChanceRAG, a No-Code Retrieval Augmented Generation Solution ChanceRAG is built around key features such as PDF processing, vector store creation, BM25 indexing, and the fusion of retrieval methods.
Read more at: https://analyticsindiamag.com/ai-news-updates/rabbitt-ai-launches-chancerag-a-no-code-retrieval-augmented-generation-solution/
Tomi Engdahl says:
Specialist ‘carbon nanotube’ AI chip built by Chinese scientists is 1st of its kind and highly energy-efficient
News
By Owen Hughes published September 4, 2024
Scientists in China have developed a tensor processing unit (TPU) that uses carbon-based transistors instead of silicon – and they say it’s extremely energy efficient.
https://www.livescience.com/technology/electronics/specialist-carbon-nanotube-ai-chip-built-by-chinese-scientists-is-1st-of-its-kind-and-1700-times-more-efficient-than-googles-version
Tomi Engdahl says:
Meet the new, most powerful open source AI model in the world: HyperWrite’s Reflection 70B
https://venturebeat.com/ai/meet-the-new-most-powerful-open-source-ai-model-in-the-world-hyperwrites-reflection-70b/
Tomi Engdahl says:
https://blog.netprofile.fi/naista-merkeista-tunnistat-tekoalyn-tuottaman-artikkelin
Tomi Engdahl says:
Kauanko sähköauton akku kestää? Nyt ilmestyi uusi jättitutkimus
Ikääntyvistä ajoakuista ilmaantuu jatkuvasti uutta tutkimustietoa.
Kauanko sähköauton akku kestää? Nyt ilmestyi uusi jättitutkimus
https://www.is.fi/autot/art-2000010710045.html
Geotab on tutkinut perusteellisesti peräti 5 000 nykyaikaisen sähköauton akustoja ja tullut siihen tulokseen, että ajoakkujen kapasiteetista häviää tätä nykyä keskimäärin 1,8 prosenttia vuodessa – mutta vain alkuvaiheessa.
Tulos on huomattava parannus aiempaan, eli Geotabin vuonna 2019 tekemään vastaavaan tutkimukseen ja tarkoittaa sitä, että uusimman sukupolven sähköautoakkujen pitäisi kestää käytössä helposti parikymmentä vuotta tai jopa vieläkin kauemmin.
Tomi Engdahl says:
https://www.geotab.com/blog/ev-battery-health/
Tomi Engdahl says:
https://bitbucket.org/product/
AI that goes beyond code generation
Atlassian infuses AI throughout the entire software lifecycle to improve developer experience and increase velocity.
AI-generated PR descriptions
Automatically generate pull request summaries based on code changes and commit messages.
Generative AI editing
Help developers write comments and PR summaries to enhance writing clarity.
AI integrations
Use your preferred AI tool to get code suggestions with context from your Bitbucket repositories or have AI take a first pass at code review.
Tomi Engdahl says:
KPMG 2024 U.S. CEO Outlook
Top CEOs see through global turbulence by betting big on AI
https://kpmg.com/us/en/articles/2024/kpmg-2024-us-ceo-outlook.html
Tomi Engdahl says:
https://etn.fi/index.php/13-news/16647-tekoaelystae-tulee-avaintekijae-kyberuhkien-torjunnassa
Tomi Engdahl says:
Tekoälystä tulee avaintekijä kyberuhkien torjunnassa
https://etn.fi/index.php/13-news/16647-tekoaelystae-tulee-avaintekijae-kyberuhkien-torjunnassa
Kriittisen infrastruktuurin suojelu on siirtymässä uuteen aikakauteen, kun tekoälystä (AI) tulee avaintekijä kyberuhkien torjunnassa. Check Point Researchin mukaan vuosina 2024 tammi–elokuussa energia- ja vesilaitoksiin kohdistui viikossa keskimäärin 1514 kyberhyökkäystä, mikä on peräti 37 % enemmän kuin edellisvuonna.
Tämä kasvu korostaa tekoälyyn perustuvien ratkaisujen merkitystä infrastruktuurin puolustuksessa. AI:n kyky käsitellä valtavia datavirtoja reaaliajassa mahdollistaa poikkeavuuksien ja uhkien havaitsemisen nopeammin ja tarkemmin kuin koskaan ennen.
Koneoppiminen mahdollistaa järjestelmien jatkuvan kehittymisen ja kyvyn pysyä kyberrikollisten edellä. Voimaverkkojen, vesihuollon ja liikenneverkkojen operaattoreille AI tarjoaa tehokkaan suojan, joka voi estää vakavia häiriöitä. Lisäksi tekoäly vapauttaa ihmisen asiantuntijat rutiinitehtävistä ja antaa heille enemmän aikaa keskittyä monimutkaisiin uhkien analysointiin ja ratkaisuihin. Vaikka tekoälyyn liittyy myös riskejä, kuten tekoälypohjaiset hyökkäykset, sen tarjoamat hyödyt infrastruktuurin suojelussa ovat merkittäviä.
Tekoäly ei pelkästään paranna uhkien havaitsemista, vaan se tehostaa myös automaatiota kriittisissä järjestelmissä. AI voi itsenäisesti analysoida hälytyksiä, yhdistää tietoja eri lähteistä ja aloittaa tarvittavat vastatoimet. Tämä nopeuttaa reaktiota kyberhyökkäyksiin ja vapauttaa ihmistiimit strategiseen suunnitteluun. Tällainen automaatio lisää kriittisen infrastruktuurin joustavuutta ja varmistaa, että järjestelmät pysyvät toimintakykyisinä myös kyberhyökkäysten aikana.
Energiasektori on yksi esimerkki siitä, miten tekoäly voi tehostaa infrastruktuuria. Älykkäissä sähköverkoissa AI ennustaa energiantarpeen vaihtelut ja optimoi energian jakelun, mikä parantaa energiatehokkuutta ja varmistaa vakaat toimitukset. Lisäksi tekoäly auttaa ennakoivassa huollossa, ennustamalla laitteistojen vikoja ja vähentäen näin käyttökatkoksia sekä ylläpitokustannuksia.
Tomi Engdahl says:
“Overbuilding things the world doesn’t have use for, or is not ready for, typically ends badly.” https://trib.al/zlqL4be
Tomi Engdahl says:
Top Goldman Sachs Stock Researcher Warns AI Bubble May Be About to Explode
byVictor Tangermann
Sep 25, 9:05 AM EDT
Michael M. Santiago/Getty Images
“Despite its expensive price tag, the technology is nowhere near where it needs to be in order to be useful.”
https://futurism.com/the-byte/goldman-sachs-ai-bubble
Ringing the Bell
The head of stock research at Goldman Sachs, Jim Covello, believes that the burgeoning AI industry could be in for a rude awakening.
As the New York Times reports, Covello closely followed the crash of the dot-com bubble months after he started working at Goldman, watching as thousands of workers were laid off.
In June, he issued a stark warning in a widely reported research paper, arguing that the billions of dollars being poured into AI companies may not see a sufficient return and that the current crop of AI tools simply aren’t good enough to bring about a worthwhile boost in productivity.
“Despite its expensive price tag, the technology is nowhere near where it needs to be in order to be useful,” Covello wrote in the report. “Overbuilding things the world doesn’t have use for, or is not ready for, typically ends badly.”
The paper marked a significant turning point, with a number of venture capitalists starting to become wary of an emerging bubble that could be set to burst.
Covello is far from alone in warning of an impending AI bubble. In a June blog post, published just days before Covello’s report, Sequoia Capital partner David Cahn argued that the entire tech industry would have to generate $600 billion per year to remain viable.
While “speculative frenzies are part of technology, and so they are not something to be afraid of,” he warned, AI tech is anything but a “get rich quick” scheme.
Other experts, like DoubleLine Capital billionaire CEO Jeffrey Gundlach, have drawn direct comparisons with the dot com bubble.
“This feels a lot like 1999,”
Tomi Engdahl says:
The Most Sophisticated AIs Are Most Likely to Lie, Worrying Research Finds
The AIs are “getting better at pretending to be knowledgeable.”
https://futurism.com/sophisticated-ai-likely-lie
Beware the smart ones: they seem to have all the answers, but can also weave the most convincing lies.
It seems that this logic also applies to large language models, which are becoming more powerful with each iteration. New research suggests that this smarter crop of AI chatbots are actually becoming less trustworthy, because they’re more likely to make up facts rather than avoiding or turning down questions they can’t answer.
The study, published in the journal Nature, examined some of the leading commercial LLMs in the industry: OpenAI’s GPT, and Meta’s LLaMA, along with an open source model called BLOOM created by the research group BigScience.
While it found that their responses are in many cases becoming more accurate, they were across the board less reliable, giving a higher proportion of wrong answers than older models did.
“They are answering almost everything these days. And that means more correct, but also more incorrect [answers],”
“That looks to me like what we would call bullshitting,” Hicks, who was not involved in the study, told Nature. “It’s getting better at pretending to be knowledgeable.”
According to the researchers, some of the biggest BS-ers were OpenAI’s GPT-4 and o1, which would answer almost any question thrown at them. But all of the studied LLMs appear to be trending this way, and for the LLaMA family of models, none of them could reach a level of 60 percent accuracy for the easiest questions, the study said.
In sum, the bigger the AI models got — in terms of parameters, training data, and other factors — the bigger the percentage of wrong answers they gave.
Still, AI models are getting better at answering more complex questions. The problem, other than their propensity for BS-ing, is that they still mess up the easy ones.
Tomi Engdahl says:
“Now, the intersection of generative AI and CGI image creation is the next wave.” https://trib.al/RTYYwNg
James Cameron, Who Warned of Skynet, Joins Board of AI Company
https://futurism.com/the-byte/james-cameron-terminator-joins-board-ai?fbclid=IwY2xjawFl2PhleHRuA2FlbQIxMQABHVJCpFaCbds1NXQOoxDAwtqhskZZE5Eb5SWoddDfT0UGsmvuxRAEkqcJXA_aem_Ud-NCSkLU8oAhzaHO04FHw
The guy behind Skynet getting into AI — what could possibly go wrong?
AI’ll Be Back
Canadian filmmaker James Cameron, the man behind the iconic film “The Terminator” — in which a rogue AI destroys most of humankind — is jumping on the AI bandwagon.
Generative AI company Stability AI, the firm behind the popular image generator Stable Diffusion, announced today that Cameron will be joining its board of directors.
“I’ve spent my career seeking out emerging technologies that push the very boundaries of what’s possible, all in the service of telling incredible stories,” Cameron said in a statement. “I was at the forefront of CGI over three decades ago, and I’ve stayed on the cutting edge since. Now, the intersection of generative AI and CGI image creation is the next wave.”
It’s a notable addition, not just because of Cameron’s sheer amount of influence in Hollywood. The use of generative AI in the entertainment industry has proven highly controversial, with actors fighting for their right to not have their likeness or voice be replaced with AI.
Tomi Engdahl says:
https://etn.fi/index.php/13-news/16658-uudet-ai-mallit-tekevaet-aelypuhelimesta-vihdoin-aelykkaeaen
Tomi Engdahl says:
MIT spinoff Liquid debuts non-transformer AI models and they’re already state-of-the-art
https://venturebeat.com/ai/mit-spinoff-liquid-debuts-non-transformer-ai-models-and-theyre-already-state-of-the-art/
Tomi Engdahl says:
Microsoft Unveils Copilot Vision AI Tool, but Highlights Security After Recall Debacle
Microsoft has unveiled a new AI-based web content analysis tool, underscoring safety and security to address potential concerns.
https://www.securityweek.com/microsoft-unveils-copilot-vision-ai-tool-but-highlights-security-after-recall-debacle/
Tomi Engdahl says:
Dylan Martin / CRN:
Accenture says it will train 30K employees on Nvidia’s AI technologies and start a new Nvidia Business Group focused on driving enterprise adoption of AI agents — The global services powerhouse says its newly formed Nvidia Business Group will focus on driving enterprise adoption …
Accenture To Train 30,000 Staff On Nvidia AI Tech In Blockbuster Deal
https://www.crn.com/news/ai/2024/accenture-to-train-30-000-staff-on-nvidia-ai-tech-in-blockbuster-deal
The global services powerhouse says its newly formed Nvidia Business Group will focus on driving enterprise adoption of what it called ‘agentic AI systems’ by taking advantage of key Nvidia software platforms that fuel consumption of GPU-accelerated data centers.
Global services powerhouse Accenture said it will train 30,000 employees on Nvidia’s full stack of AI technologies and start a new business group dedicated to the AI computing giant as part of a blockbuster deal between the two behemoths.
Accenture, No. 1 on CRN’s 2024 Solution Provider 500 list, announced the expanded partnership on Wednesday and said its newly formed Nvidia Business Group will focus on driving enterprise adoption of what it called “agentic AI systems” by taking advantage of key Nvidia software platforms that fuel consumption of GPU-accelerated data centers.
“The power of these announcements cannot be overstated,” Accenture Chief AI Officer Lan Guan said in a briefing with journalists and analysts the day before.
Called the “next frontier” of generative AI, these “agentic AI systems” involve an “army of AI agents” that work alongside human workers to “make decisions and execute with precision across even the most complex workflows,” according to Guan, a 21-year Accenture veteran. Unlike chatbots such as ChatGPT, these agents do not require prompts from humans, and they are not meant to automating pre-existing business steps.
How Accenture Plans To Use Nvidia’s Tech
Most of the work done by Accenture’s new business group will happen through its AI Refinery platform, which takes advantage of three key Nvidia software platforms:
Nvidia AI Foundry aids with the development of custom generative AI models by providing access to foundation models, frameworks and tools. Nvidia AI Enterprise offers microservices and tools for developing, optimizing and deploying AI applications. Nvidia Omniverse, on the other hand, enables development and deployment of 3-D applications.
The latter two technologies are paid offerings from Nvidia, and they require systems that are powered by the company’s popular and powerful GPUs.
Julie Sweet, chair and CEO of Accenture, said the company is “breaking significant new ground” and helping clients use generative AI as a catalyst for reinvention.”
“Accenture AI Refinery will create opportunities for companies to reimagine their processes and operations, discover new ways of working, and scale AI solutions across the enterprise to help drive continuous change and create value,” she said in a statement.
Why Accenture Sees ‘Agentic AI’ As A Big Deal
Guan called “agentic AI systems” a “game-changer in how our clients reinvent their business” and said these agents will have “self-reflecting capabilities” and a “memory engine” that will allow them to learn from mistakes.
Agents within these systems will specialize in specific tasks, whether that’s to research topics or orchestrate “multiple functions across your enterprise,” according to Guan.
For example, Accenture designed a system consisting of 14 agents that allows its marketing team to “create and run smarter campaigns faster.” The company said this is expected to reduce the manual steps taken for such tasks by 25-35 percent, decrease costs by 6 percent and accelerated completion of said tasks by 22-55 percent.
Nvidia’s Big Push For Expanded GSI Partnerships
Nvidia’s expanded partnership with Accenture is part of a concerted effort by the AI computing giant to grow its influence with global systems integrators (GSIs) and other multinational IT services companies that have significant influence within enterprises.
Among the other GSIs that have developed extended relationships with Nvidia, Infosys said a year ago that it would train and certify 50,000 employees on the chip designer’s AI hardware and software solutions for example.
Tomi Engdahl says:
Kyle Wiggers / TechCrunch:
Microsoft unveils Copilot Vision, which analyzes text and images on webpages, Think Deeper, which reasons through more complex problems, Copilot Voice, and more
Microsoft Copilot can now read your screen, think deeply, and speak aloud to you
https://techcrunch.com/2024/10/01/microsoft-copilot-can-now-read-your-screen-think-deeper-and-speak-aloud-to-you/
A week after announcing a wave of updates for its enterprise suite of Copilot AI-powered products, Microsoft is launching new Copilot capabilities on Windows for all users, including a tool that can understand and respond to questions about what’s on your screen.
Refreshed Copilot apps for iOS, Android, Windows, and the web are rolling out today, and all feature a Copilot with a more “warm” and “distinct” style, as Microsoft describes it. Microsoft is also bringing the chatbot to WhatsApp, letting users chat with Copilot via DM, similar to the experience you get with other bots on Meta’s messaging platform.
Tomi Engdahl says:
Hyung-Jin Kim / Associated Press:
A look at the prevalence of deepfake porn in South Korea, where its parliament has revised a law to make watching or possessing deepfake porn content illegal
In South Korea, deepfake porn wrecks women’s lives and deepens gender conflict
https://apnews.com/article/south-korea-deepfake-porn-women-df98e1a6793a245ac14afe8ec2366101
SEOUL, South Korea (AP) — Three years after the 30-year-old South Korean woman received a barrage of online fake images that depicted her nude, she is still being treated for trauma. She struggles to talk with men. Using a mobile phone brings back the nightmare.
“It completely trampled me, even though it wasn’t a direct physical attack on my body,” she said in a phone interview with The Associated Press. She didn’t want her name revealed because of privacy concerns.
Many other South Korean women recently have come forward to share similar stories as South Korea grapples with a deluge of non-consensual, explicit deepfake videos and images that have become much more accessible and easier to create.
It was not until last week that parliament revised a law to make watching or possessing deepfake porn content illegal.
Most suspected perpetrators in South Korea are teenage boys. Observers say the boys target female friends, relatives and acquaintances — also mostly minors — as a prank, out of curiosity or misogyny. The attacks raise serious questions about school programs but also threaten to worsen an already troubled divide between men and women.
Deepfake porn in South Korea gained attention after unconfirmed lists of schools that had victims spread online in August. Many girls and women have hastily removed photos and videos from their Instagram, Facebook and other social media accounts. Thousands of young women have staged protests demanding stronger steps against deepfake porn. Politicians, academics and activists have held forums.
Recent attention to the problem has coincided with France’s arrest in August of Pavel Durov, the founder of the messaging app Telegram, over allegations that his platform was used for illicit activities including the distribution of child sexual abuse. South Korea’s telecommunications and broadcast watchdog said Monday that Telegram has pledged to enforce a zero-tolerance policy on illegal deepfake content.
Police say they’ve detained 387 people over alleged deepfake crimes this year, more than 80% of them teenagers. Separately, the Education Ministry says about 800 students have informed authorities about intimate deepfake content involving them this year.
The U.S. cybersecurity firm Security Hero called South Korea “the country most targeted by deepfake pornography” last year. In a report, it said South Korean singers and actresses constitute more than half of the people featured in deepfake pornography worldwide.
The prevalence of deepfake porn in South Korea reflects various factors including heavy use of smart phones; an absence of comprehensive sex and human rights education in schools and inadequate social media regulations for minors as well as a “misogynic culture” and social norms that “sexually objectify women,” according to Hong Nam-hee, a research professor at the Institute for Urban Humanities at the University of Seoul.
Victims speak of intense suffering.
The 30-year-old woman interviewed by The AP said that her doctoral studies in the United States were disrupted for a year. She is receiving treatment after being diagnosed with panic disorder and post-traumatic stress disorder in 2022.
Police said they’ve detained five men for allegedly producing and spreading fake explicit contents of about 20 women, including her. The victims are all graduates from Seoul National University, the country’s top school.
The woman said the images she received on Telegram used photos she had posted on the local messaging app Kakao Talk, combined with nude photos of strangers.
The fake images were very crudely made but the woman felt deeply humiliated and shocked because dozens of people — some of whom she likely knows — were sexually harassing her with those photos.
Tomi Engdahl says:
Emma Roth / The Verge:
Google updates Lens to let users take a video and use their voice to search, showing an AI Overview and search results, rolling out to Search Labs on mobile — If you can’t capture what you want to search for with just a picture, Google Lens will now let you take a video — and even use your voice to ask about what you’re seeing.
Google Lens now lets you search with video
/ You can use your voice to ask questions about what you’re seeing, too.
https://www.theverge.com/2024/10/3/24259759/google-lens-search-video-ai-launch
Tomi Engdahl says:
Kyle Wiggers / TechCrunch:
Google rolls out AI-organized Search results in the US on mobile, showing users AI-aggregated pages of content from forums, articles, YouTube videos, and more
https://techcrunch.com/2024/10/03/google-brings-ads-to-ai-overviews-and-rolls-out-ai-organized-pages/
Emma Roth / The Verge:
Google rolls out ads in AI Overviews for when a user’s question has a “commercial angle”, and identifies the ads with a “sponsored” header, on mobile in the US — Google is rolling out ads in AI Overviews, which means you’ll now start seeing products in some of the search engine’s AI-generated summaries.
https://www.theverge.com/2024/10/3/24260637/googles-ai-overview-ads-launch
Tomi Engdahl says:
Hayden Field / CNBC:
OpenAI gets a $4B revolving line of credit from JPMorgan Chase, Citi, Goldman Sachs, and others, on top of the $6.6B in funding, bringing its liquidity to $10B+ — OpenAI has a $4 billion revolving line of credit, bringing its total liquidity to more than $10 billion, CNBC has learned.
https://www.cnbc.com/2024/10/03/openai-gets-4-billion-revolving-credit-line-on-top-of-latest-funding.html
Tomi Engdahl says:
Maxwell Zeff / TechCrunch:
OpenAI launches canvas, a ChatGPT interface with a workspace for writing and coding projects, similar to Anthropic’s Artifacts, in beta for Plus and Team users — OpenAI introduced a new way to interact with ChatGPT on Thursday: an interface it calls “canvas.”
https://techcrunch.com/2024/10/03/openai-launches-new-canvas-chatgpt-interface-tailored-to-writing-and-coding-projects/
Tomi Engdahl says:
OpenAI introduced a new way to interact with ChatGPT on Thursday: an interface it calls “canvas.” The product opens a separate window, beside the normal chat window, with a workspace for writing and coding projects. Users can generate writing or code directly in the canvas, then highlight sections of the work to have the model edit. Canvas is rolling out in beta to ChatGPT Plus and Teams users on Thursday, and Enterprise and Edu users next week.
Several consumer AI providers are converging around editable workspaces as a practical way to use generative AI. ChatGPT’s new interface offers similar features to Anthropic’s Artifacts, launched in June, and the viral coding companion, Cursor. OpenAI is racing to match competitor offerings, and launch entirely new capabilities in ChatGPT, as a means to grow its paid user base.
https://techcrunch.com/2024/10/03/openai-launches-new-canvas-chatgpt-interface-tailored-to-writing-and-coding-projects/
Tomi Engdahl says:
Even the ‘godmother of AI’ has no idea what AGI is
https://techcrunch.com/2024/10/03/even-the-godmother-of-ai-has-no-idea-what-agi-is/
Are you confused about artificial general intelligence, or AGI? It’s that thing OpenAI is obsessed with ultimately creating in a way that “benefits all of humanity.” You may want to take them seriously since they just raised $6.6 billion to get closer to that goal.
But if you’re still wondering what the heck AGI even is, you’re not alone.
In a wide ranging discussion on Thursday at Credo AI’s responsible AI leadership summit, Fei-Fei Li, a world-renowned researcher often called the “godmother of AI,” said she doesn’t know what AGI is either. At other points, Li discussed her role in the birth of modern AI, how society should protect itself against advanced AI models, and why she thinks her new unicorn startup World Labs is going to change everything.
But when asked what she thought about an “AI singularity,” Li was just as lost as the rest of us.
“I come from academic AI and have been educated in the more rigorous and evidence-based methods, so I don’t really know what all these words mean,” said Li to a packed room in San Francisco, beside a big window overlooking the Golden Gate Bridge. “I frankly don’t even know what AGI means. Like people say you know it when you see it, I guess I haven’t seen it. The truth is, I don’t spend much time thinking about these words because I think there’s so many more important things to do…”
If anyone would know what AGI is, it’s probably Fei-Fei Li.
Tomi Engdahl says:
Google Developers Blog:
Google releases Gemini 1.5 Flash-8B, a smaller and faster 1.5 Flash variant with a 50% lower price, 2x higher rate limits, and lower latency on small prompts — Today, Gemini 1.5 Flash-8B, our latest Flash variant, is production-ready and comes with: — Developers can access gemini-1.5-flash-8b …
https://developers.googleblog.com/en/gemini-15-flash-8b-is-now-generally-available-for-use/
Tomi Engdahl says:
Koodareiden hehkutetut työkalut tyrmättiin: eivät paranna tuottavuutta – virheet lisääntyvät 41 %
Aleksi Kolehmainen2.10.202422:45|päivitetty2.10.202422:45TekoälyOhjelmistokehitys
Tuore selvitys kyseenalaistaa GitHubin Copilotin kaltaisten tekoälyavusteisten koodaustyökalujen vaikutukset kehittäjien tuottavuuteen ja työuupumuksen ehkäisyyn.
https://www.tivi.fi/uutiset/koodareiden-hehkutetut-tyokalut-tyrmattiin-eivat-paranna-tuottavuutta-virheet-lisaantyvat-41-/30565064-74b7-4628-90d5-7ed0d649d8f1
Tomi Engdahl says:
Nvidia yllätti kaikki omalla avoimella LLM-mallillaan
https://etn.fi/index.php/13-news/16675-nvidia-yllaetti-kaikki-omalla-avoimella-llm-mallillaan
NVIDIA on julkaissut uuden avoimen LLM-mallin. NVLM 1.0 kilpailee johtavien multimodaalisten kielimallien, kuten OpenAI:n GPT-4:n kanssa. Se näyttää testien perusteella suoriutuvan erityisen hyvin visuaalis-kielellisistä tehtävistä eli malli osaa selittää kuvia ja eri kuvien eroja.
Toisin kuin GPT-4, jonka malliarkkitehtuurit ja datat ovat suurelta osin suljettuja, NVLM 1.0 julkaisee malliensa painot ja avaa koodinsa tutkimusyhteisön käyttöön. Tämä tarjoaa merkittävän edun tutkimuksen ja kehityksen kannalta.
NVLM 1.0:ssa on kolme arkkitehtuurivaihtoehtoa: decoder-only (NVLM-D), cross-attention (NVLM-X) ja hybridi-arkkitehtuuri (NVLM-H). Tämä eroaa GPT-4:n yhden neuroverkon mallista, joka käsittelee teksti- ja kuvasyötteitä ilman erillisiä moduuleja. NVLM tarjoaa erillisiä arkkitehtuurivaihtoehtoja, joista jokaisella on vahvuutensa erilaisissa laskentatehtävissä.
NVLM 1.0 on suunniteltu erityisesti kuvien tunnistukseen ja muihin visuaalis-kielellisiin haasteisiin. Se käyttää dynaamisia korkearesoluutioisia kuvia, mikä parantaa suorituskykyä tehtävissä, kuten tekstintunnistuksessa ja matemaattisessa päättelyssä. GPT-4 on erittäin monipuolinen, mutta se ei ole yhtä optimoitu skannattuihin tai korkearesoluutioisiin kuviin liittyvissä tehtävissä.
Tomi Engdahl says:
https://etn.fi/index.php/13-news/16674-google-kuluttaa-enemmaen-saehkoeae-kuin-moni-valtio
Tomi Engdahl says:
Tekoälyä uhkaa luhistuminen – ratkaisuja etsitään kuumeisesti
Panu Räty4.10.202406:05TekoälyData ja analytiikka
Laajojen kielimallien kehitys uhkaa hiipua, koska ihmisten tuottama koulutusaineisto on hupenemassa.
https://www.tivi.fi/uutiset/tekoalya-uhkaa-luhistuminen-ratkaisuja-etsitaan-kuumeisesti/c66a1293-355b-4426-b602-1434a1540ca7
Tomi Engdahl says:
”Oletko robotti?” Nyt se tapahtui: perinteisestä turvajärjestelmästä tuli täysin hyödytön
2.10.202413:02|päivitetty2.10.202413:02
Oletko robotti -kysymysten avulla pyritään erottamaan botit oikeaista ihmisistä. Valitettavasti botit ovat muuttumassa liian fiksuiksi.
https://www.mikrobitti.fi/uutiset/oletko-robotti-nyt-se-tapahtui-perinteisesta-turvajarjestelmasta-tuli-taysin-hyodyton/1b3c4b13-f6ac-40d2-8045-414c1b1a5701
Tomi Engdahl says:
Tekoälytaiteilija nappasi voiton ihmistaiteilijoista – valittaa nyt menettävänsä rahaa, kun muut varastavat hänen työnsä
2.10.202414:56|päivitetty2.10.202414:56
Tekoälytaiteilija haluaisi suojata teoksensa tekijänoikeudella.
https://www.mikrobitti.fi/uutiset/tekoalytaiteilija-nappasi-voiton-ihmistaiteilijoista-valittaa-nyt-menettavansa-rahaa-kun-muut-varastavat-hanen-tyonsa/30b8f393-1629-4761-825b-ad06940de9c2
Tomi Engdahl says:
Tällaisen startupin Bill Gates nyt perustaisi
Microsoftin kavereidensa kanssa perustanut Gates hyppäisi koko rahan edestä nykytrendin vietäväksi
https://www.talouselama.fi/uutiset/tallaisen-startupin-bill-gates-nyt-perustaisi/610c5245-dbfb-4972-8b50-07c8e081f414
Tomi Engdahl says:
Tekoäly muuttaa yhteiskunnan, aivot ja ihmisyyden
Edes internetin käyttöönotto ei ollut samanlainen murros kuin se, jossa elämme nyt, tietojenkäsittelytieteen professori Sasu Tarkoma arvioi.
https://www.helsinki.fi/fi/uutiset/ihmisten-teknologia/tekoaly-muuttaa-yhteiskunnan-aivot-ja-ihmisyyden
Tomi Engdahl says:
Nvidia just dropped a bombshell: Its new AI model is open, massive, and ready to rival GPT-4
https://venturebeat.com/ai/nvidia-just-dropped-a-bombshell-its-new-ai-model-is-open-massive-and-ready-to-rival-gpt-4/
Tomi Engdahl says:
Microsoft Copilot can now read your screen, think deeply, and speak aloud to you
https://techcrunch.com/2024/10/01/microsoft-copilot-can-now-read-your-screen-think-deeper-and-speak-aloud-to-you/
Tomi Engdahl says:
Nvidia pudotti todellisen uutispommin
Nvidia tarjoaa uutta tekoälymalliaan avoimesti myös muille tutkimus- ja kehityskäyttöön, mutta sen kaupallista käyttöä rajoitetaan lisenssillä.
https://www.iltalehti.fi/digiuutiset/a/e4232e37-bdd8-4825-995f-6b94a22064bd
Nvidia on lanseerannut uuden avoimen tekoälymallin, joka haastaa alan johtavat suljetut mallit, kuten esimerkiksi OpenAI:n gpt4-mallit, jotka toimivat ChatGPT:n taustalla.
Uusi kielimallien perhe tunnetaan nimellä nvlm 1.0. Se kykenee käsittelemään sekä kirjoitettua kieltä että visuaalisia syötteitä kuten kuvia ja tekstiä yhdessä.
Nvidia on julkaissut tekoälymallinsa avoimesti, mutta käyttöä rajoittaa kuitenkin sen lisenssi, joka estää kaupallisen käytön. Nvlm 1.0 -malliperhe on saatavilla creative commons -lisenssillä (cc-by-nc 4.0).
Käytännössä tämä tarkoittaa, että mallia saa vapaasti jakaa, muokata ja käyttää ei-kaupallisissa tarkoituksissa, kunhan alkuperäisille tekijöille annetaan asianmukainen tunnustus. Sen sijaan mallia ei saa käyttää esimerkiksi kaupalliseen sovellukseen ilman erillistä lupaa.
https://huggingface.co/nvidia/NVLM-D-72B
Tomi Engdahl says:
MITRE Announces AI Incident Sharing Project
MITRE’s AI Incident Sharing initiative helps organizations receive and hand out data on real-world AI incidents.
https://www.securityweek.com/mitre-announces-ai-incident-sharing-project/
Tomi Engdahl says:
Ryan Broderick / Garbage Day:
As Meta proudly unveils AI tools, Facebook users find feeds clogged with AI junk, even when Facebook is needed as a civic resource to coordinate disaster relief — Read to the end for furry opera — We’re inching closer to the Garbage Day Live election special!!! Over the next few issues, I’ll be announcing a new guest.
What feels real enough to share
Read to the end for furry opera
https://www.garbageday.email/p/what-feels-real-enough-to-share
It Doesn’t Matter If It Isn’t Real
Tomi Engdahl says:
Digiday:
TikTok launches Smart+, an AI-powered ad tool that takes on Google’s Performance Max and Meta’s Advantage+, automating creative, ad targeting, and optimization — TikTok’s cultural clout is undeniable, but the ad dollars? Not quite there — yet. — That’s where Smart+ comes in …
https://digiday.com/marketing/tiktok-joins-the-ai-driven-advertising-pack-to-compete-with-meta-for-ad-dollars/
Tomi Engdahl says:
Sara Fischer / Axios:
OpenAI signs a content deal with Hearst to use content from 40+ local newspapers and 20+ magazines in its products, after Condé Nast and Dotdash Meredith deals
https://www.axios.com/2024/10/08/openai-hearst-ai-content-deal
Tomi Engdahl says:
Mike Wheatley / SiliconANGLE:
Meta begins rolling out new AI tools that let advertisers expand the aspect ratio of their videos and generate an Instagram Reel video ad from a static image
Meta debuts new generative AI tools for creating video-based ads
https://siliconangle.com/2024/10/08/meta-debuts-new-generative-ai-tools-creating-video-based-ads/
Tomi Engdahl says:
Jess Weatherbed / The Verge:
Adobe launches a Content Authenticity web app to let creators use its Content Credentials “nutrition labels” to apply attribution and opt out of AI training
Adobe has a new tool to protect artists’ work from AI
/ The free Content Authenticity web app allows creators to apply attribution and ‘do not train’ tags to any image, video, or audio work.
https://www.theverge.com/2024/10/8/24265031/adobe-content-authenticity-web-app-ai-label-availability
Adobe is expanding its Content Credentials “nutrition labels” to make it even easier for creatives to be credited for their work, identify what is and isn’t AI online, and protect their content in the process. It’s launching a free web app that will allow users to quickly apply creator information to images, videos, and audio and even opt them out of generative AI models — for the AI developers that support it, at least.
The Content Authenticity web app can be used to widely apply attribution data to content that contains the creator’s name, website, social media pages, and more. It also provides an easier way for creatives to opt their work out of AI training en mass compared to laboriously submitting individual protections for their content to each AI provider.
Media Alert: Adobe Introduces Adobe Content Authenticity Web App to Champion Creator Protection and Attribution
https://news.adobe.com/news/2024/10/aca-announcement
shawn M davidson says:
I found this article on AI misconceptions to be very insightful. It’s so important for IT leaders to have a grounded understanding of AI, especially with all the hype surrounding it. The article highlights key issues like AI not being a magical fix for every business problem and the importance of having the right data infrastructure in place. These are critical points for anyone looking to implement AI successfully. It’s a great read for anyone interested in the practical realities of AI in business.”
Tomi Engdahl says:
https://www.convergens.fi/post/strategic-implementation-of-ai-to-build-a-competitive-advantage-by-leveraging-the-organisations-own
Tomi Engdahl says:
https://www.convergens.fi/post/strategic-implementation-of-ai-to-build-a-competitive-advantage-by-leveraging-the-organisations-own
Tomi Engdahl says:
Making A Bot $3800/Day with ChatGPT | Simple Copy-Paste Method
https://www.youtube.com/watch?v=8buLmUKtyqU
Tomi Engdahl says:
Tekoälyllä voi optimoida elämää, mutta haurauden sietämistä on hyvä vaalia
Tekoäly voi vaikuttaa siihen, millaisia ihmisiä olemme ja miten suhtaudumme toisiimme. Ihmisyyden ydintä se ei kuitenkaan voi korvata, sanovat etiikan tutkijat.
https://www.helsinki.fi/fi/uutiset/kulttuuri/tekoalylla-voi-optimoida-elamaa-mutta-haurauden-sietamista-hyva-vaalia
Tomi Engdahl says:
Apple Researchers Release Depth Pro, a One-Shot Model for Sub-Second Depth Mapping on Any Image
Calculating an accurate depth map from a single 2D image in 0.3 seconds, Depth Pro is impressive — and on GitHub now.
https://www.hackster.io/news/apple-researchers-release-depth-pro-a-one-shot-model-for-sub-second-depth-mapping-on-any-image-ef8e5e8e4683?fbclid=IwY2xjawF1fzlleHRuA2FlbQIxMQABHWPEs1yeoxFDmMXtmm18ldOJQ70TRQgWm0VNusl2OtsIECTIYTxhbd58gA_aem_gdFfy4UGUG5YA98bMEOrmQ
“We present a foundation model for zero-shot metric monocular depth estimation,” the research team explains. “Our model, Depth Pro, synthesizes high-resolution depth maps with unparalleled sharpness and high-frequency details. The predictions are metric, with absolute scale, without relying on the availability of metadata such as camera intrinsics. Extensive experiments analyze specific design choices and demonstrate that Depth Pro outperforms prior work along multiple dimensions.”
Depth mapping is handy for everything from robotic vision to blurring the background of images post-capture. Typically, it relies on being able to capture the scene from two slightly different angles — as with smartphones that have multiple rear-facing cameras, where the differences between the images on two sensors are used to calculate depth and separate the foreground from the background — or the use of a distance-measuring technology such as lidar. Depth Pro, though, requires neither of these, yet Apple claims it can turn a single two-dimensional image into an accurate depth map in well under a second.
For predicting depth, we employ two ViT encoders, a patch encoder and an image encoder.”
https://github.com/apple/ml-depth-pro?tab=readme-ov-file
Tomi Engdahl says:
Hayden Field / CNBC:
OpenAI says it has disrupted 20+ operations and deceptive networks in 2024 so far that tried to use its models, including four that had election-related content
https://www.cnbc.com/2024/10/09/openai-says-more-cyber-actors-using-its-platform-to-disrupt-elections.html