“We reproduced #DeepSeek R1-Zero in the CountDown game, and it just works
Through RL, the 3B base LM develops #SelfVerification and #search abilities all on its own
You can experience the Ahah moment yourself for < $30” — #JiayiPan
Beginning to see replication of Deepseek … “learns to allocate more thinking time to a problem by re-evaluating its approach” … this is described as the “Ah Ha Moment”.
[#]AI / #ReenforcementLearning https://github.com/Jiayi-Pan/TinyZero / https://x.com/jiayi_pirate/status/1882839370505621655?s=46 / https://youtube.com/watch?v=e659KrxxN5w
=> More informations about this toot | View the thread
“JPMorgan has reopened an underground #gold vault in New York that was mothballed in the 1990s” — FT
“US banking giant #JPMorgan plans to deliver $4bn of gold bullion weighing more than 937 tons to #NewYork this month” — TheGuardian
Listen, there's a line of #DumpTrucks northbound on the #FDR at about 70th — John McClane
=> More informations about this toot | View the thread
“Two #Brisbane #entrepreneurs whose start-up helps major companies find weaknesses in their #cybersecurity have sold out to a British private equity-backed firm for more than $100 million.
Michael #Gianarakis and Shubham #Shah, who describe themselves as #EthicalHackers, founded #Assetnote in 2018. They have since signed Atlassian, Qantas and Canva as clients, and have been profitable since they started.”
[#]Australia / #Hackers https://afr.com/technology/brisbane-ethical-hackers-bank-100m-payday-from-uk-buyer-20250130-p5l8ej
=> More informations about this toot | View the thread
🚨 #TechWeenie Alert
Headed for technofascism’: the rightwing roots of Silicon Valley
“An influential Silicon Valley publication runs a cover story lamenting the “pussification” of tech. A major tech CEO lambasts a Black civil rights leader’s calls for diversifying the tech workforce. #Technologists rage against the “PC police”.
No, this isn’t #SiliconValley in the age of Maga. It’s the tech industry of the 1990s, when observers first raised concerns about the rightwing bend of Silicon Valley and the potential for “#technofascism”. Despite the industry’s (often undeserved) reputation for #liberalism, its reactionary foundations were baked in almost from the beginning. As Silicon Valley enters a second Trump administration, the gendered roots of its original reactionary movement offer insight into today’s rightward turn.”
[#]BeccaLewis / #oligarch / #TechOligarch https://www.theguardian.com/technology/ng-interactive/2025/jan/29/silicon-valley-rightwing-technofascism
=> More informations about this toot | View the thread
“the recent #erasure and #submerging of abortion-related content, so soon after Donald Trump’s return to #power, has caused fear among some abortion rights supporters that a bigger crackdown could be on the horizon – and concerns that the #Meta founder #MarkZuckerberg’s promise to protect “#FreeExpression” may not apply to all speech.”
🚨 #TechWeenie Alert
[#]tech / #oligarch https://theguardian.com/world/2025/jan/29/abortion-pills-instagram-shadow-banning
=> More informations about this toot | View the thread
“Experts have also questioned the assumption that DeepSeek was building with 10,000 A100 #Nvidia chips, with analysts like #DylanPatel speculating that #DeepSeek needs at least 50,000 of Nvidia’s far-more powerful chips, the H100s. #Meta for instance, operates the equivalent of 600,000 Nvidia H100s.”
Courtesy of Bloomberg.
=> More informations about this toot | View the thread
“Earlier that Monday, Liang attended a closed-door business symposium in #Beijing that was hosted by Chinese #PremierLiQiang. There, experts in technology, science, education and other fields offered their opinions and suggestions for a draft government work report, according to the official Xinhua news agency. Video footage on YouTube shows Liang sitting across the table from Li and speaking, with the Chinese leader nodding attentively.
Significantly, #DeepSeek open sourced its #R1, allowing researchers and developers to freely use, modify and commercialise the model. That sent a signal that it wants to collaborate and innovate with others in the global #AI community.”
=> More informations about this toot | View the thread
“#DeepSeek’s research was funded by #HighFlyer’s R&D budget, Liang said previously. It drew computing resources from the quant fund, which had amassed 10,000 Nvidia #GPU s in 2021, prior to US bans on exports of sophisticated #Nvidia chips and other graphics processing units.”
=> More informations about this toot | View the thread
[#]Computerphile #MikePound, University of #Nottingham gives a CS level technical explanation of #DeepSeek.
[#]AI / #CompSci / #UniNottingham https://youtube.com/watch?v=gY4Z-9QlZ64
=> More informations about this toot | View the thread
“We asked #DeepSeek’s #AI questions about topics historically censored by the #GreatFirewall. Here’s how its responses compared to the free versions of #ChatGPT and Google’s #Gemini #chatbot.
‘Sorry, that’s beyond my current scope’
Unsurprisingly, DeepSeek did not provide answers to questions about certain political events. When asked the following questions, the AI assistant responded: “Sorry, that’s beyond my current scope. Let’s talk about something else.”
[#]Politics / #China / #Taiwan https://www.theguardian.com/technology/2025/jan/28/we-tried-out-deepseek-it-works-well-until-we-asked-it-about-tiananmen-square-and-taiwan
=> More informations about this toot | View the thread
“How many "R"s are in the word strawberry?”
… is the new #TuringTest
=> More informations about this toot | View the thread
“While certainly an improvement over non-CoT models in terms of math #reasoning, we're not sure we can fully trust R1 or any other model's #math skills just yet, especially when giving the model a #calculator is still faster.”
great article on #TheRegister testing #R1
[#]AI / #CoT / #DeepSeek 🧮 https://www.theregister.com/2025/01/26/deepseek_r1_ai_cot/?td=rt-3a
=> More informations about this toot | View the thread
“While some local #tech firms and #DataCentre operators have taken a hit on the #ASX, the biggest falls have been for #Uranium producers.. ‘Interestingly, the biggest impact has been on uranium stocks, off the back of fears that the adoption of #NuclearEnergy to power data centres will be delayed and diminished by the impacts of DeepSeek on the AI complex,’”
The #DeepSeek shakeout contines…
[#]Energy / #AI / #technology / #nuclear ☢️ https://www.abc.net.au/news/2025-01-28/asx-markets-business-news-live-updates/104865804
=> More informations about this toot | View the thread
“Deepseek R1 Explained by a Retired Microsoft Engineer”
Dave from “Daves’ Garage” does a good job summarising #DeepSeek R1 from a Software Engineers perspective.
[#]AI / #hardware / #DavePlumber / #software / #SoftwareEngineering
=> More informations about this toot | View the thread
“An Indiana man who was pardoned by Donald Trump for taking part in the January 6 insurrection was killed by police during a traffic stop on Sunday.
Matthew Huttle, 42, was shot by a sheriff’s deputy after allegedly resisting arrest and getting into an altercation with an officer, local news outlets in Indiana report, based on the Indiana state police’s account of the incident.”
Brass verdict.
[#]MichaelConnolley / #Bosch https://www.theguardian.com/us-news/2025/jan/27/jan-6-pardon-police-killing-matthew-huttle
=> More informations about this toot | View the thread
“Reuters is reporting that one of DeepSeek's research papers showed that it had used about 2,000 of Nvidia's #H800 chips, which were designed to comply with US export controls released in 2022.
The #US microchip export controls were designed to freeze #China's development of supercomputers used to develop nuclear weapons and artificial intelligence systems.
"#DeepSeek didn't come out of nowhere — they've been at model building for years," #RAND Corporation technology advisor Jimmy Goodrich told #Reuters.”
You’ll see a lot of BS arguments over the number & type of NV chips used. This is Politics.
The key thing to remember is the low price will increase use. The production numbers (costs) will be verified because the code is open source.
=> More informations about this toot | View the thread
“$955b meltdown: Wall Street darling suffers biggest one-day loss in history:
… The semiconductor maker (#Nvidia) is leading a broader selloff in #technology shares after #DeepSeek’s low-cost approach reignited concerns that big #US companies have poured too much money into developing artificial intelligence, since the #Chinese firm appears to provide a comparable #performance to Western #chatbots at a fraction of the #price.”
Hardly a response for an economy / investment based on reality. AI development will rumble on for decades more, with / without the “stupid money”.
[#]AI / #WallStreet https://smh.com.au/business/markets/955-billion-meltdown-wall-street-darling-suffers-biggest-one-day-loss-in-history-20250128-p5l7lt.html
=> More informations about this toot | View the thread
“#DeepSeek claims to have used fewer chips than its rivals to develop its models, making them cheaper to produce and raising questions over a multibillion-dollar #AI spending spree by US companies that has boosted markets in recent years.
The company developed bespoke #algorithms to build its models using reduced-capability H800 chips produced by Nvidia, according to a research paper published in December.
Nvidia’s most advanced chips, H100s, have been banned from export to China since September 2022 by US sanctions. #Nvidia then developed the less powerful #H800 chips for the #Chinese market, although they were also banned from export to #China last October.”
This is hilarious on two fronts. A) a better product with cheaper production processes, B) Foreign owned Co.
Making things cheaper is the key to #innovation & uptake. That’s why US #investment is spooked. No expensive hurdles.
https://www.theguardian.com/business/2025/jan/27/tech-shares-asia-europe-fall-china-ai-deepseek
=> More informations about this toot | View the thread
“It was clearly designed to get attention. I don’t intend to add to that attention because I do think that it takes away from what the day should be about – which is the amazing people who were nominated as Australians of the Year.”
— Australian PM on Australia Day, 2025
Translation, “How dare you embarrass me in front of Australian Oligarchs when I could be at my 4M AUD beach house enjoying Straya day. 🤣☺️
=> More informations about this toot | View the thread
“A string of major blazes – in 2006, 2013, 2014 and 2024 – have burned 90% of the Grampians landscape, Plumanns Pouton says. “The issue with having so many fires in such a short time frame is that plants need enough time to be able to accumulate seed again.”
[#]ClimateChange / #Victoria / #TheGrampians / #Bushfires https://www.theguardian.com/environment/2025/jan/27/victoria-grampians-fires-endangered-globe-pea-plant-rescue
=> More informations about this toot | View the thread
=> This profile with reblog | Go to peterrenshaw@ioc.exchange account This content has been proxied by September (3851b).Proxy Information
text/gemini
