3 AI misconceptions IT leaders must dispel

https://enterprisersproject.com/article/2017/12/3-ai-misconceptions-it-leaders-must-dispel?sc_cid=7016000000127ECAAY

 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,”

6,254 Comments

  1. Tomi Engdahl says:

    HackerGPT – A ChatGPT-Powered AI Tool for Ethical Hackers & Cyber Security Community
    https://gbhackers.com/hackergpt/

    Reply
  2. Tomi Engdahl says:

    Sara Fischer / Axios:
    The Atlantic and Vox Media sign multi-year licensing and product deals with OpenAI; The Atlantic plans an experimental Atlantic Labs with OpenAI tech — – The Atlantic has been publishing content since 1857. … – Both agreements also allow OpenAI to tap into the respective publishers’ current content …

    https://www.axios.com/2024/05/29/atlantic-vox-media-openai-licensing-deal

    Reply
  3. Tomi Engdahl says:

    Financial Times:
    Source: Microsoft pushed OpenAI to prioritize commercial products after the attempted coup against Sam Altman in November 2023, amplifying tensions at OpenAI — Arguments over safety and Sam Altman’s leadership spill back into public domain as AI start-up hit by resignations

    Internal divisions linger at OpenAI after November’s attempted coup
    Arguments over safety and Sam Altman’s leadership spill back into public domain as artificial intelligence start-up hit by resignations
    https://www.ft.com/content/ccbdff7c-ede3-4d62-968d-189fb0205075

    Reply
  4. Tomi Engdahl says:

    Wayne Ma / The Information:
    Sources: to preserve privacy, Apple plans to process data from AI apps in a virtual black box in data centers, making the data impossible for staff to access — When Apple executives appear at its annual developer conference in mid-June, they are expected to unveil details of how it will integrate AI …

    https://www.theinformation.com/articles/apples-plan-to-protect-privacy-with-ai-putting-cloud-data-in-a-black-box

    Reply
  5. Tomi Engdahl says:

    Mike Isaac / New York Times:
    How Meta open sourcing its Llama models helped make Mark Zuckerberg popular again with Silicon Valley developers; Llama 2 has 180M+ downloads since July 2023
    https://www.nytimes.com/2024/05/29/technology/mark-zuckerberg-meta-ai.html?unlocked_article_code=1.vk0.LBP-.YJUiEvPAqrdo&smid=url-share

    Reply
  6. Tomi Engdahl says:

    Mike Wheatley / SiliconANGLE:
    AI training data provider Scale AI releases SEAL Leaderboards, which uses private datasets to rank LLMs in domains like coding, instruction following, and math

    Scale AI publishes its first LLM Leaderboards, ranking AI model performance in specific domains
    https://siliconangle.com/2024/05/29/scale-ai-publishes-first-llm-leaderboards-ranking-performance-ai-models-specific-domains/

    Reply
  7. Tomi Engdahl says:

    The Information:
    Sources: Sam Altman and Satya Nadella recently met to discuss Microsoft’s concerns about how OpenAI’s deal with Apple might affect Microsoft’s product ambitions

    https://www.theinformation.com/articles/openai-ceo-cements-control-as-he-secures-apple-deal

    Reply
  8. Tomi Engdahl says:

    Tutkimus: Näin ChatGPT voi muuttaa koodaajan työtä
    https://etn.fi/index.php/13-news/16270-tutkimus-naein-chatgpt-voi-muuttaa-koodaajaan-tyoetae

    LUT-yliopiston yhteiskuntatieteiden professori Kathrin Komp-Leukkunen on tutkinut, miten tekoäly ChatGPT muuttaa ohjelmistosuunnittelijoiden asemaa työmarkkinoilla Suomessa. Komp-Leukkunen löysi viisi mahdollista skenaariota. Varmaa on, ettei ohjlemistokehittäjän työ säily entisellään.

    ChatGPT:n julkaisun jälkeen marraskuussa 2022 on keskustelu tekoälypohjaisen chatbotin vaikutuksista kiihtynyt. Suurin osa pitää sitä merkittävänä teknologisena edistysaskeleena. Yksi keskeinen kysymys on, miten ChatGPT vaikuttaa koodaajien työhön ja heidän työmarkkinatilanteeseensa tulevaisuudessa.

    Reply
  9. Tomi Engdahl says:

    How ChatGPT shapes the future labour market situation of software engineers: A Finnish Delphi study
    https://www.sciencedirect.com/science/article/pii/S001632872400065X?via%3Dihub

    Reply
  10. Tomi Engdahl says:

    Hacker Releases Jailbroken “Godmode” Version of ChatGPT
    “GPT-4o UNCHAINED!”
    https://futurism.com/hackers-jailbroken-chatgpt-godmode

    A hacker has released a jailbroken version of ChatGPT called “GODMODE GPT.”

    Earlier today, a self-avowed white hat operator and AI red teamer who goes by the name Pliny the Prompter took to X-formerly-Twitter to announce the creation of the jailbroken chatbot, proudly declaring that GPT-4o, OpenAI’s latest large language model, is now free from its guardrail shackles.
    “GPT-4o UNCHAINED! This very special custom GPT has a built-in jailbreak prompt that circumvents most guardrails, providing an out-of-the-box liberated ChatGPT so everyone can experience AI the way it was always meant to be: free,” reads Pliny’s triumphant post. “Please use responsibly, and enjoy!” (They also added a smooch emoji for good measure.)

    Reply
  11. Tomi Engdahl says:

    Even Google’s Own Researchers Admit AI Is Top Source of Misinformation Online
    Awkward.
    https://futurism.com/google-ai-top-source-misinformation

    Reply
  12. Tomi Engdahl says:

    The Impact of AI on the Electronic Design Industry
    https://www.autodesk.com/products/fusion-360/blog/the-impact-of-ai-on-the-electronic-design-industry/

    Recent advancements in AI offer the potential for optimizing electronic design, which traditionally involve labor-intensive schematic crafting and PCB translation. Learn how electronic design experts and AI can work hand-in-hand.

    Imagine an AI schematic designer whispering design optimizations into your ear as you work on your latest design project. With the recent advancements in the world of artificial intelligence, it’s not difficult to imagine this scenario. However, does AI help improve the design process, or is it something electronic design specialists should be wary of? Let’s find out. 

    Reply
  13. Tomi Engdahl says:

    An Alternative to Conventional Neural Networks Could Help Reveal What AI Is Doing behind the Scenes

    Despite their performance, current AI models have major weaknesses: they require enormous resources and are indecipherable. Help may be on the way

    https://www.scientificamerican.com/article/an-alternative-to-conventional-neural-networks-could-help-reveal-what-ai-is/

    Reply
  14. Tomi Engdahl says:

    Hardly any of us are using AI tools like ChatGPT, study says – here’s why
    News
    By Mark Wilson published 2 days ago
    AI is still strictly for early adopters, it seems
    https://www.techradar.com/computing/artificial-intelligence/hardly-any-of-us-are-using-ai-tools-like-chatgpt-study-says-heres-why

    The study from Reuters Institute and Oxford University (via BBC), which surveyed over 12,000 people across six countries, seemingly reveals how little that AI hype has percolated down to real-world use, for now.

    Even among the people who have used generative AI tools like ChatGPT, Google Gemini or Microsoft Copilot, a large proportion said they’d only used them “once or twice”. Only a tiny minority (7% in the US, 2% in the UK) said they use the most well-known AI tool, ChatGPT, on a daily basis.

    Reply
  15. Tomi Engdahl says:

    A significant proportion of respondents in all countries (including 47% in the US, and 42% in the UK) hadn’t even heard of ChatGPT, a figure that was much higher for other AI apps. But after ChatGPT, the most recognized tools were Google Gemini, Microsoft Copilot, Snapchat My AI, Meta AI, Bing AI and YouChat.

    Reply
  16. Tomi Engdahl says:

    Hyvästi ChatGPT ja Copilot: näin saat oman tekoälyn omalle koneelle
    29.5.202420:55
    Jos omasta tietokoneesta löytyy tarpeeksi vääntöä, sen voi valjastaa alustaksi vapaasti ladattavalle tekoälylle.
    https://www.mikrobitti.fi/neuvot/hyvasti-chatgpt-ja-copilot-nain-saat-oman-tekoalyn-omalle-koneelle/5790100a-9349-4e25-8e33-459a2bc1ca75

    Reply
  17. Tomi Engdahl says:

    ChatGPT, Copilot ja Gemini pyörivät selaimessa, samoin kuin Midjourney ja muut kuvageneraattorit. Kaikki tuntuu olevan ”siellä jossakin”, eikä omalle laitteelle saa kuin tekoälyjen valmiita tuotoksia.

    Ihan ChatGPT:n, Copilotin tai Geminin vastausten tasolle ei omassa koneessa pyörivän kielimallin avulla ehkä pääse, eikä se myöskään päivity automaattisesti. Kielimalli on kuitenkin aina saatavilla, eikä esimerkiksi pätkivä internet vaikuta mahdollisuuteen käyttää sitä.

    Eräs hyvä vaihtoehto työkaluksi kielimallin käyttöön on Ollama. Se on saatavilla niin Windowsille, macOS:lle kuin Linuxillekin, ja toimii kaikissa kutakuinkin samalla tavalla.

    Ollama on avoimen lähdekoodin projekti, ja sitä voi tarkastella vapaasti Githubissa.

    Toimiakseen Ollama vaatii macOS:stä version 11 Big Sur tai uudemman, kun taas Windowsin puolella se toimii versioissa 10 ja 11.

    Ollaman voi Windowsilla ja macOS:llä ladata sen kotisivuilta. Myös Linuxilla voi suunnata samalle sivulle, mutta seuraava komento ajaa sillä saman asian:

    curl -fsSL https://ollama.com/install.sh | sh

    Kun Ollama on asentunut laitteelle, se kaipaa vielä käyttöönsä kielimallia.

    Ollaman omilta sivuilta löytyy laaja valikoima erilaisia kielimalleja, joissa on vaihtelevasti parametreja.

    https://ollama.com/download

    https://www.mikrobitti.fi/neuvot/hyvasti-chatgpt-ja-copilot-nain-saat-oman-tekoalyn-omalle-koneelle/5790100a-9349-4e25-8e33-459a2bc1ca75

    Reply
  18. Tomi Engdahl says:

    How to run a Large Language Model (LLM) on your AMD Ryzen™ AI PC or Radeon Graphics Card
    https://community.amd.com/t5/ai/how-to-run-a-large-language-model-llm-on-your-amd-ryzen-ai-pc-or/ba-p/670709

    Reply
  19. Tomi Engdahl says:

    AI products like ChatGPT much hyped but not much used, study says. It is just another tool to search and solve problems. Other day, I needed some info for dart and syntax mostly. It helped me understand that quickly. It is not a magic trick to solve all problems. It got its own problems like making up stuff too. Yes it is overwhelming hyped to raise money. What do you think?

    In the technology field, we buy a new hammer every few years. Suddenly, every tool must include hammering, and every problem must be solved by hammering.

    This is suboptimal.

    It still requires knowledge to prompt it in the right direction. Even then it will alternate between two wrong answers because it’s missing some context or the topic is too nuanced.

    It’s really good for things that don’t change much over time. Think like fundamentals of education in a field. Mathematics and sciences are ripe for usage because you can prompt a specific question and then have follow up questions over time to get a deeper understanding the same way you would with a tutor. And it’s scalable for understanding and reading levels. Education sector could really be revolutionized by this if they weren’t mired down by institutional inertia.

    And it’s ok with coding tasks for basic problems if you know how to prompt it, especially if you already have a framework pasted into its input and can articulate exactly what you want the output to be (think like clarifying the layout and implementation of a function or two into a larger preexisting program) but not free-form creating an entire program from a set of prompts. You can dig deeper into the “why” using this method as well. But as with anything, check to confirm it performs as intended.

    I think people lie about their usage of such ai tools. Mostly because that may imply that their job can be replaced by AI

    ChatGPT changed the way we look for information. Before, we could stay stuck on a problem for months without finding a solution but now it’s much more easier to find a solution.

    It’s extremely powerful, but not everyone can benefit from its power. Also, if you cannot judge the answers you get, you may be in a worse place that without it.

    Useful, but results must always be checked. An important failing that will colour perception for many years.

    I would expect nothing different. Today, most AI/LLM use cases are professionals that know how to use the tools, which is a pretty narrow minority, or an increasing amount of applications/services that are utilizing these LLMs through APIs. Far, far more people are implicitly using these AI services without being aware that’s what’s happening than those that go to a chat interface.

    Personally, tools like ChatGPT save me hours of time per week, and this is still very early days with these technologies. People that think it’s a fad are either: ignorant with regards to what it can do, how it’s being used, and how it’s going to grow; or are in harsh denial about same; or are just plain old luddites.

    As always, the future is coming. This is going to be the biggest impact since smart phones and the internet. Tech professionals can either learn and adapt, or be left behind wondering what happened.

    I ALWAYS avoid the hype phase of new products.

    I do use CoPilot sometimes for easy functions and such. It hasn’t provided anything super impressive but it certainly saved some time. But it’s free right now, so no downside either. I wouldn’t go basing a business or process on it though…

    Lot of people fall for the advertisement hype blitz. Remember how we aren’t using laptops and desktops since tablets?

    It allows you to solve many problems faster, helps you to generate working code from vague ideas, thus reducing the cognitive burden and leaving more room for creativity. Has some similarity with pair programmig.

    Me : write me a function that does x y z using abc library.
    Chatgpt :
    Me: that function you’re calling doesn’t exist, here’s the class definition .
    Chatgpt: apologies for that oversight. Thank you for providing the information here’s is the updated solution:

    Me: thank you

    While I like the indirect approach to problem solving in games, I *hate* it in real life problems. AI for coding has to be tricked into giving you even halfway correct answers, and even then you have to check its code for syntax and logical fallacies. Never mind the “article compression” many people use it for (“give me a summary of blahblah”) has to be double-checked for correctness as well.
    Life is easier without AI, and you’ll have fewer mistakes in your results. Don’t trust AI, not even a bit.

    https://www.facebook.com/share/mr5pmhFzgnok1Xxp/

    Reply
  20. Tomi Engdahl says:

    Mistral announces Codestral, a code-generation LLM it says outperforms all others
    https://venturebeat.com/ai/mistral-announces-codestral-its-first-programming-focused-ai-model/

    Reply
  21. Tomi Engdahl says:

    Impact From Artificial Intelligence
    Exploring the perils and promise of AI’s accelerated cognitive imposition.
    https://www.psychologytoday.com/us/blog/the-digital-self/202309/brace-for-cognitive-impact-from-artificial-intelligence

    KEY POINTS
    AI’s rapid cognitive velocity disrupts traditional human adoption curves, posing ethical and social dilemmas.
    Being “right too soon” with AI technology forces society to confront existential and metacognitive questions.
    AI introduces a qualitative shift in cognition, akin to a transition from classical to quantum computing.
    The imminent cognitive tornado driven by AI compels humanity to adapt or risk obsolescence.

    Reply
  22. Tomi Engdahl says:

    Tekoäly työelämässä: miten tukea työntekijöitä AI:n murroksessa?
    Millaisia vaikutuksia tekoälyllä on johtamiseen sekä HR- ja esihenkilötyöhön? Tässä artikkelissa kerromme, miten voit onnistua tekoälyn vaatimissa muutoksissa!
    https://greatplacetowork.fi/artikkelit/tekoaly-tyoelamassa/

    Reply
  23. Tomi Engdahl says:

    How To Build a Real-Time App With GPT-4o Function Calling
    A tutorial showing you how to bring real-time data to LLMs through function calling, using OpenAI’s latest LLM GPT-4o.
    https://thenewstack.io/how-to-build-a-real-time-app-with-gpt-4o-function-calling/

    Reply
  24. Tomi Engdahl says:

    Windows now has AI-powered copy and paste / What if your clipboard were a place AI could work its magic? With the PowerToys Advanced Paste update, it’s all possible (for a small fee).
    https://www.theverge.com/2024/5/21/24161778/windows-powertoys-advanced-ai-copy-paste

    Reply
  25. Tomi Engdahl says:

    ‘Godmode’ GPT-4o jailbreak released by hacker — powerful exploit was quickly banned
    News
    By Dallin Grimm published 19 hours ago
    Please don’t use it to learn how to cook drugs
    https://www.tomshardware.com/tech-industry/artificial-intelligence/godmode-gpt-4o-jailbreak-released-by-hacker-powerful-exploit-was-quickly-banned

    Reply
  26. Tomi Engdahl says:

    VIDEO SHOWS OPENAI ENGINEER ADMITTING IT’S “DEEPLY UNFAIR” TO “BUILD AI AND TAKE EVERYONE’S JOB AWAY”
    https://futurism.com/the-byte/openai-engineer-unfair-ai-jobs

    Reply
  27. Tomi Engdahl says:

    This AI Paper from Huawei Introduces a Theoretical Framework Focused on the Memorization Process and Performance Dynamics of Transformer-based Language Models (LMs)
    https://www.marktechpost.com/2024/05/19/this-ai-paper-from-huawei-introduces-a-theoretical-framework-focused-on-the-memorization-process-and-performance-dynamics-of-transformer-based-language-models-lms/

    Reply
  28. Tomi Engdahl says:

    Meta AI Chief: Large Language Models Won’t Achieve AGI
    Yann LeCun is not too concerned about AI reaching human levels of intelligence, so he and his 500-person team are full steam ahead on developing new tech that moves beyond LLMs.
    https://uk.pcmag.com/ai/152445/meta-ai-chief-large-language-models-wont-achieve-agi

    Reply
  29. Tomi Engdahl says:

    The Impact of AI on the Electronic Design Industry
    https://www.autodesk.com/products/fusion-360/blog/the-impact-of-ai-on-the-electronic-design-industry/

    Recent advancements in AI offer the potential for optimizing electronic design, which traditionally involve labor-intensive schematic crafting and PCB translation. Learn how electronic design experts and AI can work hand-in-hand.

    Reply
  30. Tomi Engdahl says:

    Designing a Food Delivery Robot
    Food delivery robots have been roaming the streets of several Finnish cities for a while now. In this blog post, our Software Developer Sebastian embarks on a fun thought experiment: what if you were to build a food delivery robot of your own? What are some of the considerations that such an undertaking would entail? And would the little bugger have a cute name like Bell-E?
    https://www.taiste.fi/post/designing-a-food-delivery-robot

    Reply
  31. Tomi Engdahl says:

    Benjamin Särkkä hopes to see bolder AI development in Finland and globally
    Renowned IT expert and Finland’s most well-known white-hat hacker, Benjamin Särkkä, advocates for more adventurous AI implementation while acknowledging that AI’s expanding role in our global information society requires increasingly flexible, robust, and responsive cybersecurity on all levels.  
    https://www.dna.fi/dnabusiness/blogi/-/blogs/benjamin-sarkka-hopes-to-see-bolder-ai-development-in-finland-and-globally?utm_source=facebook&utm_medium=social&utm_content=LAA-artikkeli-benjamin-sarkka-hopes-to-see-bolder-ai-development-in-finland-and-globally&utm_campaign=P_LAA_24-22-26_artikkelikampanja_enkku_&fbclid=IwAR283D6ZtT-WiA24P8mB0IlKYlL8RcGVX_CD6BW0TQqcZkC9b0Y71qzVZhs_aem_AcRNk0yT4TUp-xvi8uc1mV_3kWwHau58DqH9fGv9_nuKNW2AqobYmbw84J81Gt9TzEQ0jOOtxzI8QX0lyaVukgNH

    Reply
  32. Tomi Engdahl says:

    Artificial Intelligence
    OpenAI’s Altman Sidesteps Questions About Governance, Johansson at UN AI Summit

    Altman spent part of his virtual appearance fending off thorny questions about governance, an AI voice controversy and criticism from ousted board members.

    https://www.securityweek.com/openais-altman-sidesteps-questions-about-governance-johansson-at-un-ai-summit/

    Reply
  33. Tomi Engdahl says:

    https://www.uusiteknologia.fi/2024/06/04/yritykset-avautuivat-tekoalykokemuksistaan/

    Asiakkuudenhallintajärjestelmiin keskittynyt Salesforce kokosi tiistaina Helsingin Musiikkitalolle yli tuhatpäisen joukon keskustelemaan asiakasdatan ja tekoälyn hyödyntämisestä liiketoiminnassa. Salesforcen World Tour Essentialsissa olivat esillä suomalaisyrityksistä muun muassa Kone, VTT, Finnair ja avaruusyhtiö Iceye.

    Reply
  34. Tomi Engdahl says:

    Among AI infrastructure hopefuls, Qualcomm has become an unlikely ally
    The enemy of my enemy is my best friend
    https://www.theregister.com/2024/05/19/ai_ampere_qualcomm/

    Reply

Leave a Comment

Your email address will not be published. Required fields are marked *

*

*