Kol Tregaskes
LightNews LightNews
banner
koltregaskes.bsky.social
Kol Tregaskes
@koltregaskes.bsky.social
660 followers 610 following 340 posts
AI Art Creator 🎨 | AI Video Producer 🎬🤖 | AI Music Composer 🎵 | Tech & Science News Curator🔬 | Movie & Sci-fi Enthusiast 🍿 | Geek at Heart 🤓 | #AI #Tech
Posts Replies Media Videos
koltregaskes.bsky.social
2025 AI report sees huge benefits but warns of deepfakes, data leaks, job impact. Balanced guardrails needed
https://assets.publishing.service.gov.uk/media/679a0c48a77d250007d313ee/International_AI_Safety_Report_2025_accessible_f.pdf
#AI #Safety #Innovation
assets.publishing.service.gov.uk
January 29, 2025 at 4:30 PM
koltregaskes.bsky.social
China’s DeepSeek Distils OpenAI
FT says DeepSeek used GPT outputs for cheaper AI, raising IP concerns. OpenAI blocked suspect accounts.
https://www.ft.com/content/a0dfedd1-5255-4fa9-8ccc-1fe01de87ea6
#AI #IP #Innovation
www.ft.com
January 29, 2025 at 3:30 PM
Reposted by Kol Tregaskes
opensource.block.xyz
Our newest project is taking flight: meet codename goose! 🪶

Today, we launched an open source on-machine AI Agent. It’s modular, works with your preferred LLM, and integrates seamlessly with developer tools and other software via MCP.

Developers, check it out!
block.github.io/goose/blog/2...
Introducing codename goose
codename goose is your open source AI agent, automating engineering tasks and improving productivity.
block.github.io
January 28, 2025 at 5:28 PM
koltregaskes.bsky.social
Red Teaming Reveals AI Blind Spots
• According to Microsoft, 100+ AI products tested
• Simple attacks still work
• Continuous checks urged
AI security never ends
https://arxiv.org/pdf/2501.07238
#AI #Security #RedTeaming
arxiv.org
January 25, 2025 at 8:30 PM
koltregaskes.bsky.social
OpenAI’s Stem Cell Boost

- GPT-4b micro aids reprogramming
- Retro Biosciences sees gains
- Sam Altman invests

https://www.technologyreview.com/2025/01/17/1110086/openai-has-created-an-ai-model-for-longevity-science/

#AI #StemCells #Longevity
OpenAI has created an AI model for longevity science
The company is making a foray into scientific discovery with an AI built to help manufacture stem cells.
www.technologyreview.com
January 17, 2025 at 8:02 PM
koltregaskes.bsky.social
AI Transforms Search

Generative AI enables conversational queries and direct answers, replacing keywords. Google and others race to lead, while publishers fear reduced traffic.

https://www.technologyreview.com/2025/01/06/1108679/ai-generative-search-internet-breakthroughs/

#AI #Search #Tech
AI means the end of internet search as we’ve known it
Despite fewer clicks, copyright fights, and sometimes iffy answers, AI could unlock new ways to summon all the world’s knowledge.
www.technologyreview.com
January 14, 2025 at 2:30 AM
koltregaskes.bsky.social
TikTok Slams Musk's Sale Claims

Variety reports TikTok strongly refutes Musk's claims about platform's potential US sale, calling statements "pure fiction."

variety.com/2025/digital/news/tiktok-elon-musk-sale-china-pure-fiction-1236273557/

#Tech #SocialMedia #Business
TikTok Calls Report That China Is Exploring Sale of App to Elon Musk ‘Pure Fiction’
TikTok denied a report that China is exploring a sale of the app to Elon Musk to keep TikTok operational in America amid a looming U.S. ban.
variety.com
January 14, 2025 at 1:40 AM
koltregaskes.bsky.social
Vocal Sandbox: Stanford's AI Robot

Stanford's Vocal Sandbox lets robots learn tasks via verbal commands, aiding daily life. Showcased in gift packing and animation, it targets use in bakeries and care homes.

https://hai.stanford.edu/news/making-robots-real-partners-daily-life

#AI #Robotics
Making Robots Real Partners in Daily Life
Stanford’s Vocal Sandbox brings us closer to robots that can adapt, learn, and assist in real time.
hai.stanford.edu
January 14, 2025 at 1:30 AM
koltregaskes.bsky.social
US AI Export Rules Risk Global GPU Market

Oracle warns Biden's new AI controls could cut US chip market by 80%, affecting all countries without industry input.

https://www.oracle.com/news/announcement/blog/export-control-diffusion-confusion-2025-01-05/

#AI #Technology #Trade
www.oracle.com
January 14, 2025 at 1:25 AM
koltregaskes.bsky.social
Stanford HAI predicts AI will evolve into collaborative systems tackling health, education, and finance. Concerns include rising AI-driven scams and reduced US regulations.

https://hai.stanford.edu/news/predictions-ai-2025-collaborative-agents-ai-skepticism-and-new-risks

#AI #Tech #Innovation
Predictions for AI in 2025: Collaborative Agents, AI Skepticism, and New Risks
Leading Stanford faculty offer their expectations for artificial intelligence in the new year.
hai.stanford.edu
January 14, 2025 at 12:30 AM
koltregaskes.bsky.social
Automating Remote Work May Double US GDP

Epoch AI finds automating remote work could double US GDP, driven by AI advancements and high task substitutability.

https://epoch.ai/gradient-updates/consequences-of-automating-remote-work

#AI #RemoteWork #GDP
The economic consequences of automating remote work
This Gradient Updates issue investigates the economic consequences of fully automating remote work.
epoch.ai
January 13, 2025 at 11:30 PM
koltregaskes.bsky.social
Robot production set to surge, costs dropping from $100k to $1k. Annual output may reach a billion in five years. Applications in mining, construction, and manufacturing could transform the workforce.

https://benjamintodd.substack.com/p/how-quickly-could-robots-scale-up

#Robotics #AI #Tech #Future
How quickly could robots scale up?
Some notes on robot economics.
benjamintodd.substack.com
January 13, 2025 at 10:30 PM
koltregaskes.bsky.social
NVIDIA Slams Biden's AI Rule

NVIDIA criticises Biden's 'AI Diffusion' rule, claiming it hampers AI innovation and US leadership by restricting control and stifling competition. The company urges policies fostering AI advancement.

https://blogs.nvidia.com/blog/ai-policy/

#AI #Tech #Innovation
NVIDIA Statement on the Biden Administration’s Misguided 'AI Diffusion' Rule
For decades, leadership in computing and software ecosystems has been a cornerstone of American strength and influence worldwide. The federal government has wisely refrained from dictating the…
blogs.nvidia.com
January 13, 2025 at 3:03 PM
koltregaskes.bsky.social
This is so true.

Just read Elon's posts on X and the comments appear to mostly agree with every word, every graphic, every table he posts blindly.

Critical thinking has always been vital, before social media and the latest AI revolution, and it is arguably more so now. Use these steps:
...
January 13, 2025 at 12:30 PM
koltregaskes.bsky.social
US Regulators Weigh Musk-OpenAI Dispute

FTC and DOJ echo Musk’s claims over OpenAI-Microsoft ties, raising fairness concerns, says Reuters.

https://www.reuters.com/legal/ftc-weighs-musks-lawsuit-against-openai-2025-01-10/

#AI #Antitrust #Regulation
www.reuters.com
January 13, 2025 at 11:30 AM
koltregaskes.bsky.social
Starmer's AI Plan for UK

Keir Starmer's 50-point plan to make the UK a global AI leader via education, immigration, data infrastructure, and diversity.

https://www.independent.co.uk/news/uk/politics/keir-starmer-artificial-intelligence-plan-b2677947.html

#AI #UK #Tech
Starmer’s 50-point plan for artificial intelligence revealed
New government plan involves training the next generation of AI experts at UK universities, changing immigration rules to bring in AI experts now, revising data mining rules to make the UK more…
www.independent.co.uk
January 13, 2025 at 6:39 AM
koltregaskes.bsky.social
AGI might deliver abundance, but will it also teach us how to share it?
January 13, 2025 at 4:00 AM
koltregaskes.bsky.social
In a post-AGI world, abundance won't be about having more—it’ll be about needing less
January 13, 2025 at 2:00 AM
koltregaskes.bsky.social
HOW TO PROMPT OPENAI'S O1

o1 shifts AI from chat models to a powerful report generator. Learn to write comprehensive briefs, focus on goals, and leverage o1’s strengths for efficient solutions. Explore its capabilities and limitations.

Source: https://www.latent.space/p/o1-skill-issue

#AI
o1 isn’t a chat model (and that’s the point)
How to use o1 in anger: Don’t Write Prompts; Write Briefs, Focus on Goals: describe WHAT you want, not HOW you want it, and Know what o1 does and does not do well!
www.latent.space
January 12, 2025 at 10:00 PM
koltregaskes.bsky.social
China Matches US in Quantum Computing Race

Chinese scientists reveal 105-qubit processor rivalling Google's latest, according to arXiv paper. Completes tasks that would take supercomputers billions of years.
https://arxiv.org/abs/2412.11924

#QuantumComputing #Tech #Innovation
Establishing a New Benchmark in Quantum Computational Advantage with 105-qubit Zuchongzhi 3.0 Processor
In the relentless pursuit of quantum computational advantage, we present a significant advancement with the development of Zuchongzhi 3.0. This superconducting quantum computer prototype, comprising…
arxiv.org
January 12, 2025 at 8:30 PM
koltregaskes.bsky.social
Four Core Principles LLMs Need for AGI

New research reveals LLMs require embodiment, symbol grounding, causality and memory to achieve human-level intelligence.
https://arxiv.org/pdf/2501.03151v1

#AI #AGI #Research
arxiv.org
January 12, 2025 at 7:30 PM
koltregaskes.bsky.social
AI to Replace 40% of Global Work Hours by 2028

Future Today Institute reports AI will automate 40% of work hours by 2028, driving £12.7T economic growth by 2030.
https://futuretodayinstitute.com/wp-content/uploads/2024/03/TR2024_Full-Report_FINAL_LINKED.pdf

#AI #Future #Technology
futuretodayinstitute.com
January 12, 2025 at 6:30 PM
koltregaskes.bsky.social
Big Tech Controls AI Training Data

MIT Technology Review reveals YouTube provides 70% of AI training video data, while 90% comes from Western sources, creating unprecedented power concentration.

https://www.technologyreview.com/2024/12/18/1108796/this-is-where-the-data-to-build-ai-comes-from

#AI
This is where the data to build AI comes from
New findings show how the sources of data are concentrating power in the hands of the most powerful tech companies.
www.technologyreview.com
January 12, 2025 at 5:30 PM
koltregaskes.bsky.social
AI Experts Reveal 2025 Tech Future

Fast Company reports: AI teams, $100B companies, and quantum breakthroughs predicted by 2025.
https://www.fastcompany.com/91254053/25-experts-predict-how-ai-will-change-business-and-life-in-2025

#AI #Tech #Future
www.fastcompany.com
January 12, 2025 at 4:30 PM
koltregaskes.bsky.social
AGI Will Create New Resource-Cursed Nations

LessWrong predicts AGI will function like oil within 20 years, creating resource-cursed states where AI labs become wealth extractors and human labour becomes irrelevant.
https://www.lesswrong.com/posts/Mak2kZuTq8Hpnqyzb/the-intelligence-curse

#AI
The Intelligence Curse — LessWrong
With AGI, powerful actors will lose their incentives to invest in people.
www.lesswrong.com
January 12, 2025 at 3:30 PM