#groq — Public Fediverse posts

Finally, the #Harness / Orchestrator is taking the form I desire.
See that list at the bottom? Thats some of the LLM AI engines that are connected to work on a prompt.

There are two types ORCH (Orchestrator) and WORKER, there is actual 4 of different classes of router asset.

And for the first time when #Claude hit the weekly limit, the harness did not shut down

Opencode-big-pickle #Opencode is picking up the Orchestrator role when Claude dies. The bar at the top shows which engines are used most, and surprisingly #Groq (Not Grok) is doing the heaviest lifting for the worker units (WUs)

Only Claude and Opencode are agentic capable.

TLDR: Big prompt get crank crank AFK. Ape happy.

#Agentic #Ai #LLM #Vibecode

Day 2 of rearchitecting the #Ai harness.
Multiple provider #LLM hooked up
I have also installed 3 local LLMs which will run like absolute dogshit on a 4GB VPS... but we will see if its worthwhile... because we can always pump up the server. Some #Hosting providers even offer GPUs

Ive also added a bar graph (I love graphs) up the top that shows frequency of models used. Not surprisingly #Claude is up the top... but happy to see #groq (Not the fascist grok) up there too.

Now that we can spare the compute for the organ monkey let it run for a few days and see.

#vibecoding #Harnessengineering

Telegram-бот с RAG на Cloudflare Workers: база знаний без векторов и без базы данных

Строим Telegram-бота с RAG-поиском по базе знаний — без векторных БД, без эмбеддингов, без платной инфраструктуры. Поиск по ключевым словам через Jaccard, LLM через Groq, история сессий в Cloudflare KV, деплой одной командой. Стек: TypeScript + Telegraf + Cloudflare Workers.

https://habr.com/ru/articles/1046495/

#telegrambot #cloudflare_workers #typescript #llm #jaccard #groq #telegraf #serverless #knowledgebase #rag

Как я довёл расходы на LLM до нуля: почему на бесплатных тарифах параллелизм — враг

Это продолжение первой статьи про Briefka — там я описывал самого бота и базовую архитектуру каскада LLM-провайдеров. За прошедшие 4 месяца бот органически вырос с 59 до 84 пользователей, и именно на этом масштабе бесплатный каскад начал срываться на платного провайдера. Расскажу, почему так вышло и как я вернул расходы к нулю — с цифрами и кодом. Код ниже — реальные фрагменты из боевого Briefka, слегка сокращённые для читаемости: убраны логирование и сбор статистики.

https://habr.com/ru/articles/1044546/

#llm #ratelimit #asyncio #telegrambot #groq #deepseek #fallback #circuit_breaker

How the hell is Groq raising more money?

https://www.zach.be/p/how-the-hell-is-groq-raising-more

#HackerNews #Groq #Funding #TechNews #StartupInvesting #VentureCapital

How the hell is Groq raising more money?

https://www.zach.be/p/how-the-hell-is-groq-raising-more

#HackerNews #Groq #Funding #TechNews #StartupInvesting #VentureCapital

А есть ли бесплатные API нейросетей?

Третьего дня я решил сделать лид-магнит для своего Telegram-канала. Схема такая - бот собирает у пользователя текст, обрабатывает его нейросетью, выдает что-то полезное, и в конце просит подписаться на канал в обмен на результат. Aiogram 3, Python, VPS за 150 рублей - ничего необычного. Встал первый вопрос - за что платить? Бот прототипный, аудитория на входе пока еще, собственно, не особо и понятно сколько человек. Платить $20 в месяц ради теста гипотезы - нет. Мы не ищем легких путей. Пошел разбираться, что вообще бесплатного есть.

https://habr.com/ru/articles/1041398/

#бесплатные_api #groq #groq_api #openrouter #gemini_api #telegram_бот

SwiftSlate ist so eine App, die sofort hängen bleibt.

Ein systemweiter AI-Textassistent für Android: Du tippst z. B. ?fix oder ?formal direkt im Eingabefeld – und dein Text wird sofort ersetzt. Kein Copy-Paste. Kein App-Wechsel. Unterstützt Gemini, Groq und OpenAI-kompatible Endpunkte.

Für alle, die AI auf Android wirklich im Alltag nutzen wollen, ist das richtig stark.

#Android #AI #SwiftSlate #OpenSource #Gemini #Groq #Productivity #FOSS #RawInstinctAI #OpenAI

SwiftSlate ist so eine App, die sofort hängen bleibt.

Ein systemweiter AI-Textassistent für Android: Du tippst z. B. ?fix oder ?formal direkt im Eingabefeld – und dein Text wird sofort ersetzt. Kein Copy-Paste. Kein App-Wechsel. Unterstützt Gemini, Groq und OpenAI-kompatible Endpunkte.

Für alle, die AI auf Android wirklich im Alltag nutzen wollen, ist das richtig stark.

#Android #AI #SwiftSlate #OpenSource #Gemini #Groq #Productivity #FOSS #RawInstinctAI #OpenAI

NVIDIA’s new Vera Rubin platform brings together specialized chips (Vera CPUs, Rubin GPUs, Groq LPUs, and BlueField-4 DPUs) into coordinated, rack-scale systems designed for real-time AI.

The big shift: AI isn’t just about training models anymore — it’s about orchestrating entire systems to power intelligent, autonomous agents in real time.
https://www.buysellram.com/blog/the-agentic-ai-era-how-nvidia-rubin-vera-cpu-groq-3-lpus-bluefield-4-redefine-the-inference-factory/
#NVIDIAGTC #AgenticAI #VeraRubin #DataCenter #GPU #InferenceFactory #AIInfrastructure #Groq #NVIDIA #NVLink #AIHardware #technology

https://winbuzzer.com/2026/03/17/nvidia-groq-3-lpx-non-gpu-inference-rack-gtc-2026-xcxwbn/

AI Chips: Nvidia Launches Groq 3 LPX, Its First Non-GPU Rack

#AI #AIChips #NVIDIA #Groq #JensenHuang #AIInference #NvidiaBlackwell #NvidiaGPUs #GPUs #Semiconductors #AIHardware #DataCenters #BigTech #GTC2026 #LPU

US stock markets added on Monday, closing higher in New York. The Dow finished the day at 46,946 points, up 0.8 % from the previous session. A few minutes earli... https://news.osna.fm/?p=38427 | #news #amid #fuels #groq #hormuz

Nicht mehr nur GPUs: Nvidia Groq 3 LPU erhöht Inference-Durchsatz massiv https://www.computerbase.de/news/grafikkarten/nicht-mehr-nur-gpus-nvidia-groq-3-lpu-erhoeht-inference-durchsatz-massiv.96549/ #Nvidia #Groq #GTC26

SRAM. Static RAM. The stuff used for CPU caches, including AMD's 3D chips.

There have been mumbles of CPU prices spiking like DRAM....this may be part of why.

"Companies like Cerebras, Groq, and d-Matrix are designing AI inference chips that use massive amounts of on-chip SRAM instead of relying on external DRAM (HBM), which significantly reduces latency and power consumption."

nVidia bought Groq. Amazon and Cerebras just signed a deal. Cerebras’ WSE-3 chip includes 900,000 cores and 44 gigabytes of on-chip SRAM.

Wait for it...............

#ai #dram #memory #sram #datacenters #gpu #cerebras #groq #amazon #nvidia

#Nvidia secured a non-exclusive licensing agreement with #Groq, an #AIchip startup, for $20 billion. The deal aims to bring Groq’s CEO, #JonathanRoss, on board, along with their #inferencetechnology and #intellectualproperty. This move is seen as a strategic move to counter #Google’s success with #TPUs and maintain Nvidia’s dominance in the #AIchipmarket. https://spyglass.org/nvidia-groq-deal/?eicker.news #tech #media #news

#Nvidia secured a non-exclusive licensing agreement with #Groq, an #AIchip startup, for $20 billion. The deal aims to bring Groq’s CEO, #JonathanRoss, on board, along with their #inferencetechnology and #intellectualproperty. This move is seen as a strategic move to counter #Google’s success with #TPUs and maintain Nvidia’s dominance in the #AIchipmarket. https://spyglass.org/nvidia-groq-deal/?eicker.news #tech #media #news

Nvidia otorgará inmoralidad a la tecnología del competidor de chips de IA Groq y contratará a su CEO #CEO #chips #Chips_de_IA #competidor #contratará #del #En_resumen #fusiones_y_adquisiciones #groq #licencia #Nvidia #otorgará #Tecnología #ButterWord #Spanish_News Comenta tu opinión 👇
https://butterword.com/nvidia-otorgara-inmoralidad-a-la-tecnologia-del-competidor-de-chips-de-ia-groq-y-contratara-a-su-ceo/?feed_id=61481&_unique_id=694cf0f81b521

Nvidia otorgará inmoralidad a la tecnología del competidor de chips de IA Groq y contratará a su CEO #CEO #chips #Chips_de_IA #competidor #contratará #del #En_resumen #fusiones_y_adquisiciones #groq #licencia #Nvidia #otorgará #Tecnología #ButterWord #Spanish_News Comenta tu opinión 👇
https://butterword.com/nvidia-otorgara-inmoralidad-a-la-tecnologia-del-competidor-de-chips-de-ia-groq-y-contratara-a-su-ceo/?feed_id=61481&_unique_id=694cf0f81b521

Paytm teams up with Groq to power its AI workloads on the new GroqCloud hardware. The partnership promises faster inference and tighter risk‑assessment models for its services. Curious how this could reshape fintech AI? Read on. #Paytm #Groq #GroqCloud #AIInference

🔗 https://aidailypost.com/news/paytm-partners-groq-run-ai-workloads-groqcloud-hardware

🧠 Multi-Model Support: Connect to #OpenAI, #Claude, #Gemini, #Bedrock, #Groq, #Azure, or #OpenRouter to use models like #GPT4o, #Claude3, #Llama4 and more
🔄 Session Management: Save and manage multiple conversation sessions for different projects or contexts

🧠 Multi-Model Support: Connect to #OpenAI, #Claude, #Gemini, #Bedrock, #Groq, #Azure, or #OpenRouter to use models like #GPT4o, #Claude3, #Llama4 and more
🔄 Session Management: Save and manage multiple conversation sessions for different projects or contexts

HUMAIN AI: Saudi Arabia Taps Chips from NVIDIA & Groq in AI Power Play

#AI #SaudiAI #HUMAIN #NVIDIA #Groq #AIFactories #Vision2030 #TechPartnership #AIChips #MiddleEastTech

https://winbuzzer.com/2025/05/13/humain-ai-saudi-arabia-taps-chips-from-nvidia-groq-in-ai-power-play-xcxwbn/

HUMAIN AI: Saudi Arabia Taps Chips from NVIDIA & Groq in AI Power Play

#AI #SaudiAI #HUMAIN #NVIDIA #Groq #AIFactories #Vision2030 #TechPartnership #AIChips #MiddleEastTech

https://winbuzzer.com/2025/05/13/humain-ai-saudi-arabia-taps-chips-from-nvidia-groq-in-ai-power-play-xcxwbn/

• 🧠 #g1 prototype uses #Llama3170b on #Groq to create o1-like reasoning chains

• 🔬 Improves #LLM logic through prompting strategies without additional training

• 🍓 Achieves ~70% accuracy on Strawberry problem vs 0% baseline for unprompted model

• 🔓 Open-sourced to inspire #OpenSource community to develop new #AI reasoning techniques

• 🛠️ Uses dynamic Chain of Thought, multiple methods, and alternative answer exploration

• 💻 Available on #GitHub with #Streamlit and #Gradio user interfaces for testing

https://github.com/bklieger-groq/g1

• 🧠 #g1 prototype uses #Llama3170b on #Groq to create o1-like reasoning chains

• 🔬 Improves #LLM logic through prompting strategies without additional training

• 🍓 Achieves ~70% accuracy on Strawberry problem vs 0% baseline for unprompted model

• 🔓 Open-sourced to inspire #OpenSource community to develop new #AI reasoning techniques

• 🛠️ Uses dynamic Chain of Thought, multiple methods, and alternative answer exploration

• 💻 Available on #GitHub with #Streamlit and #Gradio user interfaces for testing

https://github.com/bklieger-groq/g1

👉 Llama 3.1: l'IA open source supera i modelli IA di Anthropic e OpenAI

https://gomoot.com/llama-3-1-lia-open-source-supera-i-modelli-di-anthropic-e-openai

@elonmusk #AI #blog #Claude #gpt4 #gpt4o #gpt5 #groq #ia #Llama #llama31 #LLM #Meta #news #opensource #opus35 #perplexity #picks #PoE #sonnet35 #tech #tecnologia #token

👉 Llama 3.1: l'IA open source supera i modelli IA di Anthropic e OpenAI

https://gomoot.com/llama-3-1-lia-open-source-supera-i-modelli-di-anthropic-e-openai

@elonmusk #AI #blog #Claude #gpt4 #gpt4o #gpt5 #groq #ia #Llama #llama31 #LLM #Meta #news #opensource #opus35 #perplexity #picks #PoE #sonnet35 #tech #tecnologia #token

Do you need a SPAC therapist? - Hello and welcome back to Equity, TechCrunch’s venture capital-focused podcast, w... - http://feedproxy.google.com/~r/Techcrunch/~3/pLaVzy7VrX4/ #fundings&exits #equitypodcast #outschool #startups #coinbase #degreed #fintech #unicorn #crypto #edtech #equity #chips #slice #grab #groq #spac #ipo #smb #tc #ai