https://shift-ia.com/en/artigos/protostar-reusable-science/Shiften2026-11-08T12:17:21.857ZProject ProtoStar launches open challenges to test scientific data reusehttps://shift-ia.com/artigos/protostar-ciencia-reutilizavel/Shiftpt2026-11-08T12:17:21.857ZProject ProtoStar lança desafios para testar o reúso de dados científicoshttps://shift-ia.com/en/artigos/mistral-large4/Shiften2026-11-08T12:21:49.778ZMistral Large 4 adds text-and-image inference with three reasoning levelshttps://shift-ia.com/artigos/mistral-large4/Shiftpt2026-11-08T12:21:49.778ZMistral Large 4 adiciona interpretação de texto e imagens com três níveis de raciocíniohttps://shift-ia.com/en/artigos/typed-swift-decisions/Shiften2026-11-08T13:18:43.281884+00:00Mattt launches AnyDecisionModel to return typed decisions in Swifthttps://shift-ia.com/artigos/decisoes-tipadas-swift/Shiftpt2026-11-08T13:18:43.281884+00:00Mattt lança AnyDecisionModel para retornar decisões tipadas em Swifthttps://shift-ia.com/en/artigos/inkling-weight-release/Shiften2026-11-08T13:18:43.281884+00:00Thinking Machines Lab launches Inkling 2 Small on Tinker with closed weightshttps://shift-ia.com/artigos/inkling-pesos-pendentes/Shiftpt2026-11-08T13:18:43.281884+00:00Thinking Machines Lab lança o modelo Inkling 2 Small no Tinker com pesos fechadoshttps://shift-ia.com/en/artigos/lfm3-modality-choice/Shiften2026-11-08T13:18:43.281884+00:00Liquid AI launches six LFM3 models with different input and output modalitieshttps://shift-ia.com/artigos/lfm3-escolha-modalidade/Shiftpt2026-11-08T13:18:43.281884+00:00Liquid AI lança seis modelos LFM3 para texto, visão, áudio e decisõeshttps://shift-ia.com/en/artigos/thinkingbox-database-outcomes/Shiften2026-11-08T13:18:43.281884+00:00ThinkingBox uses database-state rewards to train Qwen3.8-27B for 50 stepshttps://shift-ia.com/artigos/thinkingbox-resultado-no-banco/Shiftpt2026-11-08T13:18:43.281884+00:00ThinkingBox usa o estado do banco para treinar Qwen3.8-27B por 50 passoshttps://shift-ia.com/en/artigos/trainium-support-lookup/Shiften2026-11-08T13:18:43.281884+00:00David Corvoysier publishes a model support lookup for AWS Trainiumhttps://shift-ia.com/artigos/trainium-consulta-compatibilidade/Shiftpt2026-11-08T13:18:43.281884+00:00David Corvoysier publica catálogo de suporte ao acelerador Trainiumhttps://shift-ia.com/en/artigos/anylanguagemodel-provider-choice/Shiften2026-11-08T14:01:47.078830+00:00Mattt releases AnyLanguageModel 1.0 for local and hosted inference in Swifthttps://shift-ia.com/artigos/anylanguagemodel-escolha-do-provedor/Shiftpt2026-11-08T14:01:47.078830+00:00Mattt lança AnyLanguageModel 1.0 para executar modelos locais e hospedados em Swifthttps://shift-ia.com/en/artigos/apertus-evaluation-records/Shiften2026-11-08T15:06:05.024191+00:00Swiss AI publishes Apertus 1.5 evaluation settings alongside multimodal resultshttps://shift-ia.com/artigos/apertus-avaliacao-reproduzivel/Shiftpt2026-11-08T15:06:05.024191+00:00Swiss AI publica condições de avaliação do Apertus 1.5, modelo que aceita texto, imagens e áudiohttps://shift-ia.com/en/artigos/autosynthdata-executable-training-tasks/Shiften2026-11-08T18:07:26.980723+00:00ServiceNow uses AutoSynthData to turn capability gaps into executable training taskshttps://shift-ia.com/artigos/autosynthdata-falhas-em-tarefas/Shiftpt2026-11-08T18:07:26.980723+00:00ServiceNow usa AutoSynthData para gerar tarefas de treinamento a partir de falhas de agenteshttps://shift-ia.com/en/artigos/adapters-that-learn-the-harness/Shiften2026-11-08T12:15:01.105ZHarness Hub releases adapters trained inside coding-agent harnesseshttps://shift-ia.com/en/artigos/agent-memory-facts-procedures-evidence/Shiften2026-11-08T12:15:01.105ZHermes separates persistent agent notes, reusable skills and conversation recordshttps://shift-ia.com/en/artigos/asr-accents-beyond-the-average/Shiften2026-11-08T12:15:01.105ZOpen ASR Leaderboard adds Egyptian and Nigerian English evaluationshttps://shift-ia.com/en/artigos/claude-code-skill-for-evidence-based-reading/Shiften2026-11-08T12:15:01.105ZClaude Code uses skills to define repeatable documentation-reading taskshttps://shift-ia.com/en/artigos/claude-hooks-before-and-after-tools/Shiften2026-11-08T12:15:01.105ZClaude Code runs hooks before and after tool executionhttps://shift-ia.com/en/artigos/claude-subagents-separate-context-not-usage/Shiften2026-11-08T12:15:01.105ZClaude Code delegates work to subagents with separate contexthttps://shift-ia.com/en/artigos/cosmos-control-meets-synthetic-video/Shiften2026-11-08T12:15:01.105ZNVIDIA adds four Cosmos checkpoints for controlled simulation videohttps://shift-ia.com/en/artigos/cuga-hermes-why-the-harness-matters/Shiften2026-11-08T12:15:01.105ZIBM compares CUGA, Hermes and Deep Agentshttps://shift-ia.com/en/artigos/cwe-bench-repository-security-repairs/Shiften2026-11-08T12:15:01.105ZCWE-bench v2 adds 100 repository security repair taskshttps://shift-ia.com/en/artigos/dharma-extraction-beyond-transcription/Shiften2026-11-08T12:15:01.105ZDharma Field Extractor converts document images into typed fieldshttps://shift-ia.com/en/artigos/granite-switch-adapter-focused-patch/Shiften2026-11-08T12:15:01.105ZIBM updates Granite Switch speech and decision componentshttps://shift-ia.com/en/artigos/hermes-gateway-browser-workspace/Shiften2026-11-08T12:15:01.105ZHermes Agent adds browser access to its Desktop workspacehttps://shift-ia.com/en/artigos/lerobot-streaming-and-live-instructions/Shiften2026-11-08T12:15:01.105ZLeRobot adds dataset streaming and live policy instructionshttps://shift-ia.com/en/artigos/mcp-setup-and-trust-boundaries/Shiften2026-11-08T12:15:01.105ZClaude Code connects external tools through MCP servershttps://shift-ia.com/en/artigos/palette-slide-design-workflow/Shiften2026-11-08T12:15:01.105ZPalette separates slide design planning from code generationhttps://shift-ia.com/en/artigos/phonon-on-device-tts-and-licensing/Shiften2026-11-08T12:15:01.105ZPhonon generates speech offline under model-specific licenseshttps://shift-ia.com/en/artigos/quantization-model-hardware-and-task/Shiften2026-11-08T12:15:01.105ZTransformers uses lower-precision weights to reduce memoryhttps://shift-ia.com/en/artigos/smolagents-code-as-an-action/Shiften2026-11-08T12:15:01.105Zsmolagents uses generated code to compose tool operationshttps://shift-ia.com/en/artigos/toyblocks-controlled-robot-learning/Shiften2026-11-08T12:15:01.105ZAi2 introduces ToyBlocks to study how robot skills transferhttps://shift-ia.com/en/artigos/transformers-mic-small-realistic-checks/Shiften2026-11-08T12:15:01.105ZTransformers MIC shrinks models for fast integration testshttps://shift-ia.com/en/artigos/benchmark-protocol-matters/Shiften2026-11-08T21:04:35.125228+00:00AISI publishes evaluation results with their experimental configurationshttps://shift-ia.com/artigos/benchmark-com-protocolo/Shiftpt2026-11-08T21:04:35.125228+00:00AISI publica resultados de avaliações com suas configurações experimentaishttps://shift-ia.com/en/artigos/falcon-asr-dialect-evidence/Shiften2026-11-08T18:07:26.980723+00:00TII launches Falcon ASR with Arabic and Emirati evaluationshttps://shift-ia.com/artigos/falcon-asr-recorte-dialetal/Shiftpt2026-11-08T18:07:26.980723+00:00TII lança Falcon ASR para transcrever fala em cinco idiomashttps://shift-ia.com/en/artigos/flux-region-control/Shiften2026-11-08T15:06:05.024191+00:00FLUX.3 adds bounding boxes to prompts for image placementhttps://shift-ia.com/artigos/flux-controle-por-regioes/Shiftpt2026-11-08T15:06:05.024191+00:00FLUX.3 adiciona caixas ao prompt para posicionar imagenshttps://shift-ia.com/en/artigos/holo4-across-screens-and-apis/Shiften2026-11-08T21:04:35.125228+00:00Hcompany launches Holo4 models that combine screens, code and tool callshttps://shift-ia.com/artigos/holo4-entre-telas-e-apis/Shiftpt2026-11-08T21:04:35.125228+00:00Hcompany lança Holo4, modelos que combinam telas, código e ferramentashttps://shift-ia.com/en/artigos/lfm-local-tool-execution/Shiften2026-11-08T15:06:05.024191+00:00Liquid AI adds local agent support with LFM2.5-24B-A2Bhttps://shift-ia.com/artigos/lfm-agente-local-ferramentas/Shiftpt2026-11-08T15:06:05.024191+00:00Liquid AI lança LFM2.5 para agentes que executam tarefas locaishttps://shift-ia.com/en/artigos/ml-intern-training-acceptance/Shiften2026-11-08T18:07:26.980723+00:00ML Intern trains models with budget approval and initial checkshttps://shift-ia.com/artigos/ml-intern-treino-com-criterio/Shiftpt2026-11-08T18:07:26.980723+00:00ML Intern treina modelos após pedir aval para gastos e checar pesoshttps://shift-ia.com/en/artigos/muse-spark-video-sampling/Shiften2026-11-08T15:06:05.024191+00:00Muse Spark Open samples video frames before processing inputhttps://shift-ia.com/artigos/muse-spark-video-amostragem/Shiftpt2026-11-08T15:06:05.024191+00:00Muse Spark Open seleciona quadros antes de processar vídeoshttps://shift-ia.com/en/artigos/omlx-funded-maintenance/Shiften2026-11-08T21:04:35.125228+00:00Hugging Face hires Jun Kim to maintain oMLXhttps://shift-ia.com/artigos/omlx-manutencao-financiada/Shiftpt2026-11-08T21:04:35.125228+00:00Hugging Face contrata Jun Kim para manter o oMLXhttps://shift-ia.com/en/artigos/tts-metrics-and-listening/Shiften2026-11-08T18:07:26.980723+00:00Open TTS Leaderboard measures speech accuracy and playback delayhttps://shift-ia.com/artigos/tts-metricas-e-escuta/Shiftpt2026-11-08T18:07:26.980723+00:00Open TTS Leaderboard compara precisão da fala e espera por áudiohttps://shift-ia.com/en/artigos/provenanceguard-claim-attribution/Shiften2026-11-08T18:07:26.980723+00:00ProvenanceGuard checks claims against their cited tool sourceshttps://shift-ia.com/artigos/provenanceguard-origem-da-afirmacao/Shiftpt2026-11-08T18:07:26.980723+00:00ProvenanceGuard verifica se afirmações citam a fonte corretahttps://shift-ia.com/artigos/adapters-modelo-aprende-harness/Shiftpt2026-11-08T12:15:01.105ZHarness Hub lança adaptadores treinados para agentes de códigohttps://shift-ia.com/artigos/asr-sotaques-fora-da-media/Shiftpt2026-11-08T12:15:01.105ZOpen ASR Leaderboard adiciona avaliações de transcrição com sotaques egípcio e nigerianohttps://shift-ia.com/artigos/claude-code-skill-leitura-verificavel/Shiftpt2026-11-08T12:15:01.105ZClaude Code usa skills para definir tarefas recorrentes de leitura de documentaçãohttps://shift-ia.com/artigos/claude-hooks-antes-depois-da-acao/Shiftpt2026-11-08T12:15:01.105ZClaude Code executa hooks antes e depois das ferramentashttps://shift-ia.com/artigos/claude-subagentes-contexto-nao-gratis/Shiftpt2026-11-08T12:15:01.105ZClaude Code delega tarefas a subagentes com contexto separadohttps://shift-ia.com/artigos/cosmos-simulacao-controle-e-imagem/Shiftpt2026-11-08T12:15:01.105ZNVIDIA adiciona quatro modelos Cosmos para gerar vídeo de simulação controladohttps://shift-ia.com/artigos/cuga-hermes-peso-do-harness/Shiftpt2026-11-08T12:15:01.105ZIBM compara três sistemas de execução de agenteshttps://shift-ia.com/artigos/cwe-bench-correcao-seguranca-repos/Shiftpt2026-11-08T12:15:01.105ZCWE-bench v2 publica 100 tarefas para avaliar correções de segurançahttps://shift-ia.com/artigos/dharma-extrair-campos-nao-so-transcrever/Shiftpt2026-11-08T12:15:01.105ZDharma Field Extractor converte imagens de documentos em campos estruturadoshttps://shift-ia.com/artigos/granite-switch-patch-nos-adapters/Shiftpt2026-11-08T12:15:01.105ZIBM atualiza fala e decisão no modelo modular Granite Switchhttps://shift-ia.com/artigos/hermes-gateway-desktop-no-navegador/Shiftpt2026-11-08T12:15:01.105ZHermes Agent adiciona acesso pelo navegador ao ambiente de trabalho do agentehttps://shift-ia.com/artigos/lerobot-dados-instrucoes-e-execucao/Shiftpt2026-11-08T12:15:01.105ZLeRobot adiciona treinamento sem download integral e instruções para robôs em execuçãohttps://shift-ia.com/artigos/mcp-conectar-sem-confundir-confianca/Shiftpt2026-11-08T12:15:01.105ZClaude Code conecta ferramentas externas por servidores MCPhttps://shift-ia.com/artigos/memoria-agente-fatos-procedimentos-evidencia/Shiftpt2026-11-08T12:15:01.105ZHermes separa notas persistentes, procedimentos reutilizáveis e histórico de conversashttps://shift-ia.com/artigos/palette-slides-com-linguagem-visual/Shiftpt2026-11-08T12:15:01.105ZPalette separa planejamento de slides e geração de códigohttps://shift-ia.com/artigos/phonon-voz-local-licencas/Shiftpt2026-11-08T12:15:01.105ZPhonon gera voz sem conexão com licenças distintas por modelohttps://shift-ia.com/artigos/quantizacao-memoria-hardware-e-tarefa/Shiftpt2026-11-08T12:15:01.105ZTransformers usa quantização para reduzir memória dos modeloshttps://shift-ia.com/artigos/smolagents-codigo-como-acao/Shiftpt2026-11-08T12:15:01.105Zsmolagents usa código gerado para combinar ações de ferramentashttps://shift-ia.com/artigos/toyblocks-mundos-controlados/Shiftpt2026-11-08T12:15:01.105ZAi2 apresenta ToyBlocks para estudar habilidades de robôshttps://shift-ia.com/artigos/transformers-mic-verificacoes-pequenas-reais/Shiftpt2026-11-08T12:15:01.105ZTransformers MIC reduz modelos para testar sua integraçãohttps://shift-ia.com/en/artigos/pytorch-mps-metal-kernels/Shiften2026-11-08T14:01:47.078830+00:00PyTorch 2.15 adds Metal kernels and fixes image corruptionhttps://shift-ia.com/artigos/pytorch-mps-kernels-metal/Shiftpt2026-11-08T14:01:47.078830+00:00PyTorch 2.15 adiciona kernels Metal e corrige imagens no Machttps://shift-ia.com/en/artigos/beam-memory-beyond-active-experts/Shiften2026-11-08T15:06:05.024191+00:00Beam activates selected experts but keeps all weights in memoryhttps://shift-ia.com/artigos/beam-memoria-alem-dos-experts/Shiftpt2026-11-08T15:06:05.024191+00:00Beam usa especialistas selecionados e mantém todos os pesoshttps://shift-ia.com/en/artigos/rl-environments-on-the-hub/Shiften2026-11-08T21:04:35.125228+00:00Hugging Face Hub adds discovery for reinforcement-learning taskshttps://shift-ia.com/artigos/ambientes-rl-no-hub/Shiftpt2026-11-08T21:04:35.125228+00:00Hugging Face Hub adiciona catálogo de tarefas de aprendizado por reforçohttps://shift-ia.com/en/artigos/torchneuron-operation-trace/Shiften2026-11-08T14:01:47.078830+00:00Amazon publishes TorchNeuron profiling steps for Trainiumhttps://shift-ia.com/artigos/torchneuron-rastro-da-operacao/Shiftpt2026-11-08T14:01:47.078830+00:00Amazon publica passos para medir TorchNeuron no Trainiumhttps://shift-ia.com/en/artigos/torchtpu-native-device/Shiften2026-11-08T14:01:47.078830+00:00TorchTPU adds native PyTorch execution on Google acceleratorshttps://shift-ia.com/artigos/torchtpu-dispositivo-nativo/Shiftpt2026-11-08T14:01:47.078830+00:00TorchTPU adiciona execução nativa do PyTorch em chips Googlehttps://shift-ia.com/en/artigos/gguf-inside-transformers/Shiften2026-11-08T21:04:35.125228+00:00Transformers adds ggml kernels to run quantized GGUF modelshttps://shift-ia.com/artigos/gguf-dentro-do-transformers/Shiftpt2026-11-08T21:04:35.125228+00:00Transformers adiciona operações do ggml para executar modelos quantizadoshttps://shift-ia.com/en/artigos/vocabx-vocabulary-cost/Shiften2026-11-08T14:01:47.078830+00:00VocabX compresses grading output while keeping scores separatehttps://shift-ia.com/artigos/vocabx-custo-do-vocabulario/Shiftpt2026-11-08T14:01:47.078830+00:00VocabX comprime respostas de correção e mantém notas separadas