OpenAI Upgrades Operator with Advanced o3 Model
OpenAI is enhancing Operator, its autonomous AI agent capable of browsing the web and executing tasks within a cloud-hosted virtual machine, by replacing the GPT-4o-based model with a more advanced version based on the new o3 model. Known for superior performance in reasoning and math, o3 surpasses its predecessor across key benchmarks. While the Operator API will continue using GPT-4o, the upgraded "o3 Operator" has been fine-tuned with additional safety datasets to better handle decisions around confirmations and refusals.
OpenAI የተባለው ድርጅት ኦፕሬተር የተሰኘውን፣ በራሱ የሚንቀሳቀስና ድረ-ገጾችን መቃኘት እንዲሁም በክላውድ በተስተናገደ ቨርቹዋል ማሽን ውስጥ ተግባራትን ማከናወን የሚችለውን ሰው ሰራሽ አስተውሎት (AI) ወኪሉን አቅም እያሳደገ ነው። ይህንንም እያደረገ ያለው አሁን ያለውን በGPT-4o ላይ የተመሰረተ ሞዴል በአዲሱና የላቀ የማመዛዘንና የሒሳብ ችሎታ እንዳለው በሚነገርለት የo3 ሞዴል በመተካት ነው።አዲሱ o3 ሞዴል በቀዳሚው ሞዴል ላይ በተለያዩ ቁልፍ የመመዘኛ መስፈርቶች የተሻለ አፈጻጸም እንዳስመዘገበ ተገልጿል።
@webthreeth
OpenAI is enhancing Operator, its autonomous AI agent capable of browsing the web and executing tasks within a cloud-hosted virtual machine, by replacing the GPT-4o-based model with a more advanced version based on the new o3 model. Known for superior performance in reasoning and math, o3 surpasses its predecessor across key benchmarks. While the Operator API will continue using GPT-4o, the upgraded "o3 Operator" has been fine-tuned with additional safety datasets to better handle decisions around confirmations and refusals.
OpenAI የተባለው ድርጅት ኦፕሬተር የተሰኘውን፣ በራሱ የሚንቀሳቀስና ድረ-ገጾችን መቃኘት እንዲሁም በክላውድ በተስተናገደ ቨርቹዋል ማሽን ውስጥ ተግባራትን ማከናወን የሚችለውን ሰው ሰራሽ አስተውሎት (AI) ወኪሉን አቅም እያሳደገ ነው። ይህንንም እያደረገ ያለው አሁን ያለውን በGPT-4o ላይ የተመሰረተ ሞዴል በአዲሱና የላቀ የማመዛዘንና የሒሳብ ችሎታ እንዳለው በሚነገርለት የo3 ሞዴል በመተካት ነው።አዲሱ o3 ሞዴል በቀዳሚው ሞዴል ላይ በተለያዩ ቁልፍ የመመዘኛ መስፈርቶች የተሻለ አፈጻጸም እንዳስመዘገበ ተገልጿል።
@webthreeth
A Five-Level Roadmap to Transformational AI
Bloomberg’s reporting on OpenAI’s vision outlines a clear, five-step progression for artificial intelligence: beginning with basic conversational chatbots (Level 1), advancing to human-level reasoners capable of complex problem solving (Level 2), then to autonomous agents that can execute tasks (Level 3). At Level 4, AI systems evolve into innovators that actively contribute to new inventions, and ultimately mature into full-scale organizational units (Level 5) able to perform the collective functions of an entire company.
የብሉምበርግ ዘገባ ስለ ኦፕንኤአይ (OpenAI) ራዕይ ሲገልጽ፣ ለአርቴፊሻል ኢንተለጀንስ (AI) ግልጽ የሆነ ባለ አምስት ደረጃ የእድገት ሂደት እንዳለው ይዘረዝራል፦ ይህም መሠረታዊ የውይይት ቻትቦቶች (ደረጃ 1) በመጀመር፣ ውስብስብ ችግሮችን መፍታት ወደሚችሉ የሰው ልጅ የማመዛዘን ደረጃ (ደረጃ 2) በማደግ፣ ከዚያም ተግባራትን በራሳቸው ማከናወን ወደሚችሉ ራሳቸውን የቻሉ ወኪሎች (ደረጃ 3) ይሸጋገራል። በደረጃ 4 ላይ፣ የ AI ስርዓቶች አዳዲስ ግኝቶችን በንቃት በማበርከት ወደ ፈጠራ ፈጣሪዎችነት ይለወጣሉ፣ እና በመጨረሻም የድርጅትን ሙሉ ተግባራት ማከናወን ወደሚችሉ ሙሉ ድርጅታዊ ክፍሎች (ደረጃ 5) ያድጋሉ።
@webthreeth
Bloomberg’s reporting on OpenAI’s vision outlines a clear, five-step progression for artificial intelligence: beginning with basic conversational chatbots (Level 1), advancing to human-level reasoners capable of complex problem solving (Level 2), then to autonomous agents that can execute tasks (Level 3). At Level 4, AI systems evolve into innovators that actively contribute to new inventions, and ultimately mature into full-scale organizational units (Level 5) able to perform the collective functions of an entire company.
የብሉምበርግ ዘገባ ስለ ኦፕንኤአይ (OpenAI) ራዕይ ሲገልጽ፣ ለአርቴፊሻል ኢንተለጀንስ (AI) ግልጽ የሆነ ባለ አምስት ደረጃ የእድገት ሂደት እንዳለው ይዘረዝራል፦ ይህም መሠረታዊ የውይይት ቻትቦቶች (ደረጃ 1) በመጀመር፣ ውስብስብ ችግሮችን መፍታት ወደሚችሉ የሰው ልጅ የማመዛዘን ደረጃ (ደረጃ 2) በማደግ፣ ከዚያም ተግባራትን በራሳቸው ማከናወን ወደሚችሉ ራሳቸውን የቻሉ ወኪሎች (ደረጃ 3) ይሸጋገራል። በደረጃ 4 ላይ፣ የ AI ስርዓቶች አዳዲስ ግኝቶችን በንቃት በማበርከት ወደ ፈጠራ ፈጣሪዎችነት ይለወጣሉ፣ እና በመጨረሻም የድርጅትን ሙሉ ተግባራት ማከናወን ወደሚችሉ ሙሉ ድርጅታዊ ክፍሎች (ደረጃ 5) ያድጋሉ።
@webthreeth
❤1
Mistral AI rolls out faster, more powerful Agents inside Le Chat
Mistral AI's update to Le Chat introduces "Agents," which are AI entities capable of utilizing all tools and connectors available in a standard chat, potentially making them the fastest agents in the industry due to their enhanced functionality and integration capabilities.
ሚስትራል ኤአይ ለ"ሌ ቻት" ያደረገው ዝመና "ኤጀንቶች" የሚባሉ አርቴፊሻል ኢንተለጀንስ አካላትን አስተዋውቋል፡፡ እነዚህ ኤጀንቶች በመደበኛ ቻት ውስጥ ያሉትን ሁሉንም መሳሪያዎች እና ማገናኛዎች የመጠቀም ችሎታ ያላቸው ሲሆን፣ በተሻሻሉ ተግባራቶቻቸው እና ውህደት ችሎታቸው ምክንያት በኢንዱስትሪው ውስጥ ፈጣን ወኪሎች ሊሆኑ ይችላሉ።
@webthreeth
Mistral AI's update to Le Chat introduces "Agents," which are AI entities capable of utilizing all tools and connectors available in a standard chat, potentially making them the fastest agents in the industry due to their enhanced functionality and integration capabilities.
ሚስትራል ኤአይ ለ"ሌ ቻት" ያደረገው ዝመና "ኤጀንቶች" የሚባሉ አርቴፊሻል ኢንተለጀንስ አካላትን አስተዋውቋል፡፡ እነዚህ ኤጀንቶች በመደበኛ ቻት ውስጥ ያሉትን ሁሉንም መሳሪያዎች እና ማገናኛዎች የመጠቀም ችሎታ ያላቸው ሲሆን፣ በተሻሻሉ ተግባራቶቻቸው እና ውህደት ችሎታቸው ምክንያት በኢንዱስትሪው ውስጥ ፈጣን ወኪሎች ሊሆኑ ይችላሉ።
@webthreeth
Web 3.0 Ethiopia - DeFi & AI
A Five-Level Roadmap to Transformational AI Bloomberg’s reporting on OpenAI’s vision outlines a clear, five-step progression for artificial intelligence: beginning with basic conversational chatbots (Level 1), advancing to human-level reasoners capable of…
Officially, Level 3 has entered for AI Evolution 🥋
Productivity Tip: Operator + Replit Agent Outperforms Replit Agent Alone
In a striking demonstration of autonomous development, Peter Gostev, Head of AI at Moonpig, handed full control of his Replit account to OpenAI’s Operator and let it work in tandem with the Replit Agent—without any manual intervention. The result was a fully built, tested, and deployed game, all achieved through AI-to-AI collaboration. What’s more compelling is that this setup performed significantly better than using the Replit Agent alone.
While the Replit Agent can build apps from prompts, it still depends on users for testing, iteration, and deployment decisions. In contrast, Operator orchestrated the full loop: prompt generation, testing feedback, iterative improvements, and final deployment. This layered agentic workflow showcases a leap in productivity and signals what’s possible when you give smart agents end-to-end autonomy.
@webthreeth
In a striking demonstration of autonomous development, Peter Gostev, Head of AI at Moonpig, handed full control of his Replit account to OpenAI’s Operator and let it work in tandem with the Replit Agent—without any manual intervention. The result was a fully built, tested, and deployed game, all achieved through AI-to-AI collaboration. What’s more compelling is that this setup performed significantly better than using the Replit Agent alone.
While the Replit Agent can build apps from prompts, it still depends on users for testing, iteration, and deployment decisions. In contrast, Operator orchestrated the full loop: prompt generation, testing feedback, iterative improvements, and final deployment. This layered agentic workflow showcases a leap in productivity and signals what’s possible when you give smart agents end-to-end autonomy.
@webthreeth
UAE Leads the Way with Free ChatGPT Plus Access for All
The United Arab Emirates (UAE) has emerged as the first country to provide free access to ChatGPT Plus, OpenAI’s premium AI chatbot, for all its citizens and residents. This groundbreaking move is part of a strategic collaboration with OpenAI and is aligned with the UAE's visionary Stargate UAE initiative. The program seeks to establish the UAE as a global leader in artificial intelligence by embedding advanced AI capabilities across key sectors such as education, healthcare, and government services.
ዩናይትድ አረብ ኤሚሬትስ (UAE) የOpenAI ፕሪሚየም AI ቻትቦት የሆነውን ChatGPT Plusን ለሁሉም ዜጎቿ እና ነዋሪዎቿ በነፃ ተደራሽ ያደረገች የመጀመሪያዋ ሀገር ሆና ብቅ ብላለች። ይህ እመርታዊ እርምጃ ከOpenAI ጋር የተደረገው ስትራቴጂካዊ ትብብር አካል ሲሆን፣ ከUAEው አርቆ አሳቢ የStargate UAE እቅድ ጋር የተጣጣመ ነው። ፕሮግራሙ እንደ ትምህርት፣ ጤና አጠባበቅ እና የመንግስት አገልግሎቶች ባሉ ቁልፍ ዘርፎች ውስጥ የላቁ የAI ችሎታዎችን በማካተት ዩናይትድ አረብ ኤሚሬትስን በአርቴፊሻል ኢንተለጀንስ የዓለም መሪ እንድትሆን ለማስቻል ይፈልጋል።
@webthreeth
The United Arab Emirates (UAE) has emerged as the first country to provide free access to ChatGPT Plus, OpenAI’s premium AI chatbot, for all its citizens and residents. This groundbreaking move is part of a strategic collaboration with OpenAI and is aligned with the UAE's visionary Stargate UAE initiative. The program seeks to establish the UAE as a global leader in artificial intelligence by embedding advanced AI capabilities across key sectors such as education, healthcare, and government services.
ዩናይትድ አረብ ኤሚሬትስ (UAE) የOpenAI ፕሪሚየም AI ቻትቦት የሆነውን ChatGPT Plusን ለሁሉም ዜጎቿ እና ነዋሪዎቿ በነፃ ተደራሽ ያደረገች የመጀመሪያዋ ሀገር ሆና ብቅ ብላለች። ይህ እመርታዊ እርምጃ ከOpenAI ጋር የተደረገው ስትራቴጂካዊ ትብብር አካል ሲሆን፣ ከUAEው አርቆ አሳቢ የStargate UAE እቅድ ጋር የተጣጣመ ነው። ፕሮግራሙ እንደ ትምህርት፣ ጤና አጠባበቅ እና የመንግስት አገልግሎቶች ባሉ ቁልፍ ዘርፎች ውስጥ የላቁ የAI ችሎታዎችን በማካተት ዩናይትድ አረብ ኤሚሬትስን በአርቴፊሻል ኢንተለጀንስ የዓለም መሪ እንድትሆን ለማስቻል ይፈልጋል።
@webthreeth
Web 3.0 Ethiopia - DeFi & AI
UAE Leads the Way with Free ChatGPT Plus Access for All The United Arab Emirates (UAE) has emerged as the first country to provide free access to ChatGPT Plus, OpenAI’s premium AI chatbot, for all its citizens and residents. This groundbreaking move is part…
Admin Thought
The agentic capabilities of the OpenAI o3 model sometimes look superhuman to me as a user. Here is a slight hope that the Government of Ethiopia follows suit like the Five Millions Coder Initiative so that this superhuman tool is available for everyone.
The agentic capabilities of the OpenAI o3 model sometimes look superhuman to me as a user. Here is a slight hope that the Government of Ethiopia follows suit like the Five Millions Coder Initiative so that this superhuman tool is available for everyone.
👍3
Google Launches AI Edge Gallery for Offline Multimodal AI on Mobile
Google has introduced the AI Edge Gallery, an open-source application that allows users to run AI models like Gemma 3 directly on Android devices, with iOS support on the horizon. Designed for offline use, the app enables multimodal AI processing without the need for an internet connection.
At its core is Gemma 3—a lightweight, mobile-optimized model based on Google's Gemini 2.0—capable of handling tasks such as image analysis and conversational AI. Key features like "Ask Image," "Prompt Lab," and "AI Chat" showcase the app’s versatility in delivering advanced AI experiences on the edge.
ቅርብ ጊዜ ጉግል የAI Edge Gallery የተሰኘ ክፍት ምንጭ የሆነ መተግበሪያ አስተዋውቋል። ይህ መተግበሪያ ተጠቃሚዎች እንደ ጄማ 3 ያሉ የAI ሞዴሎችን በቀጥታ በአንድሮይድ መሳሪያዎች ላይ እንዲያሄዱ የሚያስችል ሲሆን፣ ለiOS ድጋፍም በቅርቡ ይቀርባል። ከመስመር ውጪ ጥቅም ላይ እንዲውል ታስቦ የተሰራው ይህ መተግበሪያ የበይነመረብ ግንኙነት ሳያስፈልገው ባለብዙ-ሞዳል AI ሂደትን ያስችላል።
@webthreeth
Google has introduced the AI Edge Gallery, an open-source application that allows users to run AI models like Gemma 3 directly on Android devices, with iOS support on the horizon. Designed for offline use, the app enables multimodal AI processing without the need for an internet connection.
At its core is Gemma 3—a lightweight, mobile-optimized model based on Google's Gemini 2.0—capable of handling tasks such as image analysis and conversational AI. Key features like "Ask Image," "Prompt Lab," and "AI Chat" showcase the app’s versatility in delivering advanced AI experiences on the edge.
ቅርብ ጊዜ ጉግል የAI Edge Gallery የተሰኘ ክፍት ምንጭ የሆነ መተግበሪያ አስተዋውቋል። ይህ መተግበሪያ ተጠቃሚዎች እንደ ጄማ 3 ያሉ የAI ሞዴሎችን በቀጥታ በአንድሮይድ መሳሪያዎች ላይ እንዲያሄዱ የሚያስችል ሲሆን፣ ለiOS ድጋፍም በቅርቡ ይቀርባል። ከመስመር ውጪ ጥቅም ላይ እንዲውል ታስቦ የተሰራው ይህ መተግበሪያ የበይነመረብ ግንኙነት ሳያስፈልገው ባለብዙ-ሞዳል AI ሂደትን ያስችላል።
@webthreeth
👍1
Chapa unveils “Bilicho,” an AI chatbot that turns its dense payment-system manuals into plain answers for developers
Chapa’s in-house lab, Chapa AI Research (ChAIR), has unveiled Bilicho, a smart assistant described in a new technical report by ChAIR engineers Anwar Misbah and Israel Goytom. Built on GPT-3.5 Turbo with a retrieval layer that pulls the right snippet before it writes, Bilicho lifts answer accuracy from 70 % to 95 % and cuts developers’ search time by roughly 40 %.
የቻፓ የራሱ የጥናትና ምርምር ተቋም የሆነው ቻፓ ኤአይ ሪሰርች (Chapa AI Research - ChAIR) በድርጅቱ መሐንዲሶች አንዋር ሚስባህ እና እስራኤል ጎይቶም የተዘጋጀ የቴክኒክ ሪፖርት ይፋ አድርጓል። ሪፖርቱ “ቢሊቾ” የተሰኘ አዲስ አስተዋይ ረዳት (smart assistant) የተገለጸበት ሲሆን፣ ይህ ረዳት GPT-3.5 Turbo መሰረት አድርጎ የተገነባና ከመጻፉ በፊት ትክክለኛውን መረጃ ክፍልፋይ (snippet) መልሶ በማምጣት (retrieval layer) የመልስ ትክክለኛነትን ከ70% ወደ 95% ከፍ የሚያደርግ መሆኑ ተገልጿል። ቻፓ እንዳስታወቀው፣ ቢሊቾ የዴቨሎፐሮችን የፍለጋ ጊዜ በግምት በ40% በመቀነስ የኢንቴግሬሽን ስህተቶችን የሚቀንስ ይሆናል።
@webthreeth
Chapa’s in-house lab, Chapa AI Research (ChAIR), has unveiled Bilicho, a smart assistant described in a new technical report by ChAIR engineers Anwar Misbah and Israel Goytom. Built on GPT-3.5 Turbo with a retrieval layer that pulls the right snippet before it writes, Bilicho lifts answer accuracy from 70 % to 95 % and cuts developers’ search time by roughly 40 %.
የቻፓ የራሱ የጥናትና ምርምር ተቋም የሆነው ቻፓ ኤአይ ሪሰርች (Chapa AI Research - ChAIR) በድርጅቱ መሐንዲሶች አንዋር ሚስባህ እና እስራኤል ጎይቶም የተዘጋጀ የቴክኒክ ሪፖርት ይፋ አድርጓል። ሪፖርቱ “ቢሊቾ” የተሰኘ አዲስ አስተዋይ ረዳት (smart assistant) የተገለጸበት ሲሆን፣ ይህ ረዳት GPT-3.5 Turbo መሰረት አድርጎ የተገነባና ከመጻፉ በፊት ትክክለኛውን መረጃ ክፍልፋይ (snippet) መልሶ በማምጣት (retrieval layer) የመልስ ትክክለኛነትን ከ70% ወደ 95% ከፍ የሚያደርግ መሆኑ ተገልጿል። ቻፓ እንዳስታወቀው፣ ቢሊቾ የዴቨሎፐሮችን የፍለጋ ጊዜ በግምት በ40% በመቀነስ የኢንቴግሬሽን ስህተቶችን የሚቀንስ ይሆናል።
@webthreeth
👍1
DeepSeek Releases R1-0528 Update with AI Enhancements
DeepSeek announced the release of its R1-0528 update, bringing notable improvements to its large reasoning model, DeepSeek-R1. The update introduces enhanced deep reasoning capabilities, more natural and coherent language generation, and support for extended thinking sessions lasting up to 60 minutes.
DeepSeek የተባለው ኩባንያ R1-0528 የተሰኘውን አዲስ ማሻሻያ በታላቁ የማመዛዘን ሞዴሉ DeepSeek-R1 ላይ ማውጣቱን አስታውቋል። ይህ ማሻሻያ የተሻሻሉ ጥልቅ የማመዛዘን ችሎታዎችን፣ ይበልጥ ተፈጥሯዊና ወጥ የሆነ የቋንቋ አገላለጽን እንዲሁም እስከ 60 ደቂቃ የሚደርሱ የተራዘሙ የማሰብ ክፍለ ጊዜዎችን ይዞ መጥቷል።
@webthreeth
DeepSeek announced the release of its R1-0528 update, bringing notable improvements to its large reasoning model, DeepSeek-R1. The update introduces enhanced deep reasoning capabilities, more natural and coherent language generation, and support for extended thinking sessions lasting up to 60 minutes.
DeepSeek የተባለው ኩባንያ R1-0528 የተሰኘውን አዲስ ማሻሻያ በታላቁ የማመዛዘን ሞዴሉ DeepSeek-R1 ላይ ማውጣቱን አስታውቋል። ይህ ማሻሻያ የተሻሻሉ ጥልቅ የማመዛዘን ችሎታዎችን፣ ይበልጥ ተፈጥሯዊና ወጥ የሆነ የቋንቋ አገላለጽን እንዲሁም እስከ 60 ደቂቃ የሚደርሱ የተራዘሙ የማሰብ ክፍለ ጊዜዎችን ይዞ መጥቷል።
@webthreeth
Safaricom Commits $500 Million to Bolster East Africa's AI Infrastructure
Safaricom plans to invest $500 million over the next three years to build out artificial-intelligence infrastructure across East Africa, funding new data centers, edge-computing capacity, and digital-skills programs that will let local developers create AI solutions for agriculture, healthcare, and financial services. The company says Africa must define its own AI trajectory—shifting from passive consumption to active creation—by harmonizing data and digital laws, fostering inclusive policies, and spurring private-sector innovation.
ግንባር ቀደም የቴሌኮሙኒኬሽን አገልግሎት ሰጪ የሆነው ሳፋሪኮም በቀጣዮቹ ሦስት ዓመታት የምሥራቅ አፍሪካን ሰው ሰራሽ የማሰብ ችሎታ (AI) መሠረተ ልማት ለማልማት የ500 ሚሊዮን ዶላር ግዙፍ የኢንቨስትመንት ዕቅድ ይፋ አደረገ። ይህ ፕሮጀክት አዳዲስ የመረጃ ማዕከላትን (data center) ለመገንባት፣ የዳርቻ የኮምፒውቲንግ (edge-compute) አቅምን ለማሳደግ እና የዲጂታል ክህሎት ፕሮግራሞችን ለማካሄድ የገንዘብ ድጋፍ ያደርጋል። እነዚህ ጥረቶች የአገር ውስጥ አልሚዎች እንደ ግብርና፣ ጤና አጠባበቅ እና የገንዘብ አገልግሎቶች ባሉ ቁልፍ ዘርፎች ላይ ያተኮሩ የፈጠራ AI መፍትሔዎችን እንዲፈጥሩ ለማስቻል ያለሙ ናቸው።
@webthreeth
Safaricom plans to invest $500 million over the next three years to build out artificial-intelligence infrastructure across East Africa, funding new data centers, edge-computing capacity, and digital-skills programs that will let local developers create AI solutions for agriculture, healthcare, and financial services. The company says Africa must define its own AI trajectory—shifting from passive consumption to active creation—by harmonizing data and digital laws, fostering inclusive policies, and spurring private-sector innovation.
ግንባር ቀደም የቴሌኮሙኒኬሽን አገልግሎት ሰጪ የሆነው ሳፋሪኮም በቀጣዮቹ ሦስት ዓመታት የምሥራቅ አፍሪካን ሰው ሰራሽ የማሰብ ችሎታ (AI) መሠረተ ልማት ለማልማት የ500 ሚሊዮን ዶላር ግዙፍ የኢንቨስትመንት ዕቅድ ይፋ አደረገ። ይህ ፕሮጀክት አዳዲስ የመረጃ ማዕከላትን (data center) ለመገንባት፣ የዳርቻ የኮምፒውቲንግ (edge-compute) አቅምን ለማሳደግ እና የዲጂታል ክህሎት ፕሮግራሞችን ለማካሄድ የገንዘብ ድጋፍ ያደርጋል። እነዚህ ጥረቶች የአገር ውስጥ አልሚዎች እንደ ግብርና፣ ጤና አጠባበቅ እና የገንዘብ አገልግሎቶች ባሉ ቁልፍ ዘርፎች ላይ ያተኮሩ የፈጠራ AI መፍትሔዎችን እንዲፈጥሩ ለማስቻል ያለሙ ናቸው።
@webthreeth
❤3👍1
Benchmark Results of DeepSeek R1 – Open-Source Giant Strikes Again
DeepSeek-R1-0528 across six diverse evaluation datasets compared to major models like OpenAI-o3, Gemini-2.5-Pro-0506, and Qwen3-235B. DeepSeek-R1-0528 leads in mathematical reasoning tasks, ranking first in both AIME 2024 (91.4%) and AIME 2025 (87.5%), demonstrating exceptional strength in competitive math domains.
It also performs competitively in general question answering (GPQA Diamond) and coding (LiveCodeBench), although it slightly trails OpenAI-o3 in those benchmarks. While OpenAI-o3 dominates in agentic and reasoning-heavy tasks like Aider and Humanity’s Last Exam, DeepSeek-R1-0528 remains a close contender and significantly outperforms other open-source models, including its earlier version.
@webthreeth
DeepSeek-R1-0528 across six diverse evaluation datasets compared to major models like OpenAI-o3, Gemini-2.5-Pro-0506, and Qwen3-235B. DeepSeek-R1-0528 leads in mathematical reasoning tasks, ranking first in both AIME 2024 (91.4%) and AIME 2025 (87.5%), demonstrating exceptional strength in competitive math domains.
It also performs competitively in general question answering (GPQA Diamond) and coding (LiveCodeBench), although it slightly trails OpenAI-o3 in those benchmarks. While OpenAI-o3 dominates in agentic and reasoning-heavy tasks like Aider and Humanity’s Last Exam, DeepSeek-R1-0528 remains a close contender and significantly outperforms other open-source models, including its earlier version.
@webthreeth
❤1
Web 3.0 Ethiopia - DeFi & AI
Benchmark Results of DeepSeek R1 – Open-Source Giant Strikes Again DeepSeek-R1-0528 across six diverse evaluation datasets compared to major models like OpenAI-o3, Gemini-2.5-Pro-0506, and Qwen3-235B. DeepSeek-R1-0528 leads in mathematical reasoning tasks…
I tried the model last night, and although it isn't an O3 Level for me, it is still very good and mimics it in a very good way. I wouldn't know if it had OpenAI UI.
Should I rewrite it in Amharic? The above post?
Should I rewrite it in Amharic? The above post?
👍2
https://www.linkedin.com/feed/update/urn:li:activity:7333940258956357632/
I did a long-deep dive on what makes DeepSeek results very impressive in the lens of Organizational Efficiency plus Dynamic Capability Theory. You can read it, as it is around a page long in LinkedIn. You will love the POV of the article very much. Should I continue doing long contents like this one as well?
I did a long-deep dive on what makes DeepSeek results very impressive in the lens of Organizational Efficiency plus Dynamic Capability Theory. You can read it, as it is around a page long in LinkedIn. You will love the POV of the article very much. Should I continue doing long contents like this one as well?
Linkedin
The newly released model of DeepSeek AI has become one of the leading models through several independent benchmarks conducted by…
The newly released model of DeepSeek AI has become one of the leading models through several independent benchmarks conducted by analysts. I believe, as a student of strategy and organization, the lessons DeepSeek AI provides for strategy and business school…
❤1👍1
AI Summer Camp 2025: Free Hands-On AI Program for Grade 1–12 Students
The Ethiopian Artificial Intelligence Institute is now accepting applications for its 100 % free AI Summer Camp 2025, an immersive program where Grade 1–12 students with a strong interest in artificial intelligence can deepen their knowledge through workshops, hands-on projects, and a focused curriculum spanning machine learning, robotics, and natural-language processing. ; interested participants who are willing to engage fully in all activities can apply at https://forms.gle/yFThcRq3bSMcFEKU6 and request further details via Telegram @AI_Summer_Camp_2025 or email eaiisummercamp@gmail.com.
የኢትዮጵያ ሰዉ ሰራሽ የማሰብ ችሎታ ኢንስቲትዩት ከ1ኛ እስከ 12ኛ ክፍል ላሉና ለሰዉ ሰራሽ የማሰብ ችሎታ ከፍተኛ ፍላጎት ላላቸዉ ተማሪዎች በ2025 ሙሉ በሙሉ ነፃ የሆነ የክረምት ካምፕ ስልጠና እንደሚሰጥ አሳወቀ፡፡ ይህ የተግባር ልምምድ፥ ወርክሾፖችና በማሽን መማር፣ በሮቦቲክስ፣ በተፈጥሮ ቋንቋ ሂደት እና በሌሎችም ጉዳዮች ላይ ያተኮረ ሥርዓተ ትምህርት ተካቷል፡፡
Source - @TikvahethMagazine
@webthreeth
The Ethiopian Artificial Intelligence Institute is now accepting applications for its 100 % free AI Summer Camp 2025, an immersive program where Grade 1–12 students with a strong interest in artificial intelligence can deepen their knowledge through workshops, hands-on projects, and a focused curriculum spanning machine learning, robotics, and natural-language processing. ; interested participants who are willing to engage fully in all activities can apply at https://forms.gle/yFThcRq3bSMcFEKU6 and request further details via Telegram @AI_Summer_Camp_2025 or email eaiisummercamp@gmail.com.
የኢትዮጵያ ሰዉ ሰራሽ የማሰብ ችሎታ ኢንስቲትዩት ከ1ኛ እስከ 12ኛ ክፍል ላሉና ለሰዉ ሰራሽ የማሰብ ችሎታ ከፍተኛ ፍላጎት ላላቸዉ ተማሪዎች በ2025 ሙሉ በሙሉ ነፃ የሆነ የክረምት ካምፕ ስልጠና እንደሚሰጥ አሳወቀ፡፡ ይህ የተግባር ልምምድ፥ ወርክሾፖችና በማሽን መማር፣ በሮቦቲክስ፣ በተፈጥሮ ቋንቋ ሂደት እና በሌሎችም ጉዳዮች ላይ ያተኮረ ሥርዓተ ትምህርት ተካቷል፡፡
Source - @TikvahethMagazine
@webthreeth
❤1👍1
Factory AI's Droids: Pioneering AI Agents in Software Development or Just Another Hype?
Factory AI’s launch of “Droids” positions them as a bold new entrant in the race to redefine software development, claiming to be the world’s first AI agents capable of autonomously handling tasks across the development lifecycle.
While this move echoes a broader industry push—seen in efforts like Microsoft’s Azure AI Foundry and open-source projects like OpenHands—Factory’s emphasis on context-aware, end-to-end execution raises the question: is this truly the future of AI-driven coding, or just the latest wave of tech hype?
@webthreeth
Factory AI’s launch of “Droids” positions them as a bold new entrant in the race to redefine software development, claiming to be the world’s first AI agents capable of autonomously handling tasks across the development lifecycle.
While this move echoes a broader industry push—seen in efforts like Microsoft’s Azure AI Foundry and open-source projects like OpenHands—Factory’s emphasis on context-aware, end-to-end execution raises the question: is this truly the future of AI-driven coding, or just the latest wave of tech hype?
@webthreeth
❤1
Jules Expands Free Tier as Agentic Coding Enters Mainstream
Google's coding agent, Jules, has expanded its free daily task limit to 60 as part of its public beta launched on May 20, 2025, according to a company blog post. In addition to the increased task count, users can now run 5 concurrent tasks and access 5 codecasts per day. Jules operates autonomously in a secure Google Cloud environment, managing tasks like writing tests and fixing bugs, while integrating seamlessly with GitHub.
የጉግል የኮዲንግ ወኪል የሆነው “ጁልስ” (Jules)፤ ግንቦት 12 ቀን 2017 ዓ.ም. ይፋ ባደረገው የሙከራ (public beta) ሥሪቱ፤ በነጻ የሚሰጠውን ዕለታዊ የተግባር ብዛት ወደ 60 ከፍ ማድረጉን የኩባንያው የብሎግ መግለጫ አስታውቋል። ከተግባር ብዛት መጨመር በተጨማሪ ተጠቃሚዎች በአንድ ጊዜ አምስት የተለያዩ ተግባራትን (concurrent tasks) ማከናወን የሚችሉ ሲሆን፤ በቀን አምስት የኮድካስት (codecasts) አገልግሎቶችንም ማግኘት ይችላሉ።
@webthreeth
Google's coding agent, Jules, has expanded its free daily task limit to 60 as part of its public beta launched on May 20, 2025, according to a company blog post. In addition to the increased task count, users can now run 5 concurrent tasks and access 5 codecasts per day. Jules operates autonomously in a secure Google Cloud environment, managing tasks like writing tests and fixing bugs, while integrating seamlessly with GitHub.
የጉግል የኮዲንግ ወኪል የሆነው “ጁልስ” (Jules)፤ ግንቦት 12 ቀን 2017 ዓ.ም. ይፋ ባደረገው የሙከራ (public beta) ሥሪቱ፤ በነጻ የሚሰጠውን ዕለታዊ የተግባር ብዛት ወደ 60 ከፍ ማድረጉን የኩባንያው የብሎግ መግለጫ አስታውቋል። ከተግባር ብዛት መጨመር በተጨማሪ ተጠቃሚዎች በአንድ ጊዜ አምስት የተለያዩ ተግባራትን (concurrent tasks) ማከናወን የሚችሉ ሲሆን፤ በቀን አምስት የኮድካስት (codecasts) አገልግሎቶችንም ማግኘት ይችላሉ።
@webthreeth
AI-Powered Tigrigna Content Moderation Platform
Sustainable Technology Solutions has launched an AI-based system that evaluates Tigrigna-language social-media posts and flags them as true or false, aiming to curb misinformation, hate speech, and discrimination across online communities. Developed with grant support from the Canada Development Agency, ELIDA, and Search for Common Ground, the platform provides real-time content verification to promote healthier digital discourse among diverse groups.
Sustainable Technology Solutions የተሰኘ ተቋም በትግርኛ ቋንቋ የማኅበራዊ ሚዲያ መረጃዎችን እውነት ወይም ሐሰት መሆናቸውን የሚገመግም በአርቴፊሻል ኢንተለጀንስ (AI) ላይ የተመሠረተ ሥርዓት ይፋ አድርጓል። ይህ ሥርዓት በመስመር ላይ ማኅበረሰቦች ውስጥ የተሳሳቱ መረጃዎችን፣ የጥላቻ ንግግሮችን እና መድልዎን ለመግታት ያለመ ነው። ይህ መድረክ ከካናዳ የልማት ኤጀንሲ (Canada Development Agency)፣ ከኤሊዳ (ELIDA) እና ከሰርች ፎር ኮመን ግራውንድ (Search for Common Ground) በተገኘ የገንዘብ ድጋፍ የተዘጋጀ ሲሆን፣ በተለያዩ ቡድኖች መካከል ጤናማ የዲጂታል ውይይትን ለማበረታታት የእውነተኛ ጊዜ የይዘት ማረጋገጫ አገልግሎት ይሰጣል።
@webthreeth
Sustainable Technology Solutions has launched an AI-based system that evaluates Tigrigna-language social-media posts and flags them as true or false, aiming to curb misinformation, hate speech, and discrimination across online communities. Developed with grant support from the Canada Development Agency, ELIDA, and Search for Common Ground, the platform provides real-time content verification to promote healthier digital discourse among diverse groups.
Sustainable Technology Solutions የተሰኘ ተቋም በትግርኛ ቋንቋ የማኅበራዊ ሚዲያ መረጃዎችን እውነት ወይም ሐሰት መሆናቸውን የሚገመግም በአርቴፊሻል ኢንተለጀንስ (AI) ላይ የተመሠረተ ሥርዓት ይፋ አድርጓል። ይህ ሥርዓት በመስመር ላይ ማኅበረሰቦች ውስጥ የተሳሳቱ መረጃዎችን፣ የጥላቻ ንግግሮችን እና መድልዎን ለመግታት ያለመ ነው። ይህ መድረክ ከካናዳ የልማት ኤጀንሲ (Canada Development Agency)፣ ከኤሊዳ (ELIDA) እና ከሰርች ፎር ኮመን ግራውንድ (Search for Common Ground) በተገኘ የገንዘብ ድጋፍ የተዘጋጀ ሲሆን፣ በተለያዩ ቡድኖች መካከል ጤናማ የዲጂታል ውይይትን ለማበረታታት የእውነተኛ ጊዜ የይዘት ማረጋገጫ አገልግሎት ይሰጣል።
@webthreeth
Anthropic’s Niche-Generative Strategy for Capturing Generative-AI Market Share
Anthropic looks to be executing a focused, market-segmentation play rather than a blanket supremacy push: with Claude 4 Opus, the company concentrates R&D on human-centric use cases—adaptive reasoning, creative front-end development, and agentic terminal workflows—where real-world versatility and “taste” drive customer value.
Its first-place finishes on SimpleBench, ARC-AGI-2, Web-Dev Arena, and Terminal Bench underscore this strength, even as it willingly cedes math-heavy and highly specialized coding benchmarks to broader, maximize-everything rivals. By doubling down on differentiated capabilities that resonate with enterprise productivity and end-user experience, Anthropic is positioning itself to capture meaningful generative-AI market share without having to top every leaderboard.
Source - Peter Gostev
@webthreeth
Anthropic looks to be executing a focused, market-segmentation play rather than a blanket supremacy push: with Claude 4 Opus, the company concentrates R&D on human-centric use cases—adaptive reasoning, creative front-end development, and agentic terminal workflows—where real-world versatility and “taste” drive customer value.
Its first-place finishes on SimpleBench, ARC-AGI-2, Web-Dev Arena, and Terminal Bench underscore this strength, even as it willingly cedes math-heavy and highly specialized coding benchmarks to broader, maximize-everything rivals. By doubling down on differentiated capabilities that resonate with enterprise productivity and end-user experience, Anthropic is positioning itself to capture meaningful generative-AI market share without having to top every leaderboard.
Source - Peter Gostev
@webthreeth
Google’s New “Learn AI Skills” Portal Empowers Millions with Practical AI Education
Google has opened a new “Learn AI Skills” portal that bundles free and low-cost offerings—starting with the five-module, self-paced Google AI Essentials course and Google Cloud’s beginner-to-advanced generative-AI learning paths—to give anyone a clear route to building practical, workplace-ready AI knowledge. The program awards shareable Google certificates on completion.
ግል "የ AI ክህሎቶችን ይማሩ" (Learn AI Skills) በሚል ስያሜ አዲስ የመስመር ላይ መድረክ ይፋ ማድረጉን አስታውቋል። ይህ መድረክ ማንኛውም ሰው ተግባራዊና ለሥራ ቦታ ዝግጁ የሚያደርግ የሰው ሰራሽ የማሰብ ችሎታ (AI) ዕውቀት እንዲገነባ ግልጽ የሆነ መንገድ ለማቅረብ ያለመ ነው። ከእነዚህም መካከል ባለ አምስት ሞጁል ሲሆን በራስ መርሃ ግብር የሚጠናቀቀው "የጉግል ኤ አይ መሰረታዊያን" (Google AI Essentials) ኮርስ፣ እንዲሁም የጉግል ክላውድ ከጀማሪ እስከ ከፍተኛ ደረጃ ያሉ የ"ጀነሬቲቭ ኤ አይ" (generative AI) የመማሪያ መስመሮች ይገኙበታል።
https://ai.google/learn-ai-skills/
https://cloud.google.com/blog/topics/training-certifications/new-generative-ai-trainings-from-google-cloud
@webthreeth
Google has opened a new “Learn AI Skills” portal that bundles free and low-cost offerings—starting with the five-module, self-paced Google AI Essentials course and Google Cloud’s beginner-to-advanced generative-AI learning paths—to give anyone a clear route to building practical, workplace-ready AI knowledge. The program awards shareable Google certificates on completion.
ግል "የ AI ክህሎቶችን ይማሩ" (Learn AI Skills) በሚል ስያሜ አዲስ የመስመር ላይ መድረክ ይፋ ማድረጉን አስታውቋል። ይህ መድረክ ማንኛውም ሰው ተግባራዊና ለሥራ ቦታ ዝግጁ የሚያደርግ የሰው ሰራሽ የማሰብ ችሎታ (AI) ዕውቀት እንዲገነባ ግልጽ የሆነ መንገድ ለማቅረብ ያለመ ነው። ከእነዚህም መካከል ባለ አምስት ሞጁል ሲሆን በራስ መርሃ ግብር የሚጠናቀቀው "የጉግል ኤ አይ መሰረታዊያን" (Google AI Essentials) ኮርስ፣ እንዲሁም የጉግል ክላውድ ከጀማሪ እስከ ከፍተኛ ደረጃ ያሉ የ"ጀነሬቲቭ ኤ አይ" (generative AI) የመማሪያ መስመሮች ይገኙበታል።
https://ai.google/learn-ai-skills/
https://cloud.google.com/blog/topics/training-certifications/new-generative-ai-trainings-from-google-cloud
@webthreeth
❤1
Exa’s Purpose-Built Engine Outperforms Google on Key Search-Quality Benchmarks
Exa is a new search engine designed for AI tools by using AI rather than human web-browsing, and recent head-to-head tests show it beats Google on answer quality. In side-by-side evaluations, Exa’s responses were graded 3–12 percentage points higher than Google’s by large-language-model judges. Exa achieves this edge by controlling every step itself—collecting the web pages, reading them with its own AI models, and ranking results for clarity.
ኤክሳ (Exa) የተባለ አዲስ የፍለጋ ሞተር ይፋ ሆኗል። ይህ የፍለጋ ሞተር የተዘጋጀው ሰዎች እንደወትሮው ድረ-ገጾችን እንዲያፈላልጉበት ሳይሆን፣ ለሰው ሰራሽ የማሰብ ችሎታ (AI) መሣሪያዎች መረጃ እንዲያቀርብ ነው። በቅርቡ በተደረጉ የአቻ ለአቻ ንጽጽራዊ ሙከራዎችም ኤክሳ ከጉግል የተሻለ የምላሽ ጥራት እንዳለው ተረጋግጧል። ኤክሳ ይህን የላቀ ውጤት ያስመዘገበው እያንዳንዱን የሥራ ሂደት ማለትም ድረ-ገጾችን መሰብሰብ፣ የተሰበሰቡትን መረጃዎች በራሱ የሰው ሰራሽ የማሰብ ችሎታ ሞዴሎች ማንበብ እና ውጤቶችንም በግልጽነታቸው መሰረት ደረጃ በማውጣት በራሱ ቁጥጥር ስር በማድረጉ እንደሆነ ተገልጿል።
https://exa.ai/search (I tried it, and it seems a very good one as well).
@webthreeth
Exa is a new search engine designed for AI tools by using AI rather than human web-browsing, and recent head-to-head tests show it beats Google on answer quality. In side-by-side evaluations, Exa’s responses were graded 3–12 percentage points higher than Google’s by large-language-model judges. Exa achieves this edge by controlling every step itself—collecting the web pages, reading them with its own AI models, and ranking results for clarity.
ኤክሳ (Exa) የተባለ አዲስ የፍለጋ ሞተር ይፋ ሆኗል። ይህ የፍለጋ ሞተር የተዘጋጀው ሰዎች እንደወትሮው ድረ-ገጾችን እንዲያፈላልጉበት ሳይሆን፣ ለሰው ሰራሽ የማሰብ ችሎታ (AI) መሣሪያዎች መረጃ እንዲያቀርብ ነው። በቅርቡ በተደረጉ የአቻ ለአቻ ንጽጽራዊ ሙከራዎችም ኤክሳ ከጉግል የተሻለ የምላሽ ጥራት እንዳለው ተረጋግጧል። ኤክሳ ይህን የላቀ ውጤት ያስመዘገበው እያንዳንዱን የሥራ ሂደት ማለትም ድረ-ገጾችን መሰብሰብ፣ የተሰበሰቡትን መረጃዎች በራሱ የሰው ሰራሽ የማሰብ ችሎታ ሞዴሎች ማንበብ እና ውጤቶችንም በግልጽነታቸው መሰረት ደረጃ በማውጣት በራሱ ቁጥጥር ስር በማድረጉ እንደሆነ ተገልጿል።
https://exa.ai/search (I tried it, and it seems a very good one as well).
@webthreeth
👌2