The GPT-5.x Pro series has remained the best models for hard technical problems since they launched. There is some parallel model magic going on that is not well-explained. Anthropic has never had an equivalent, and Google basically seemed to give up on their Deep Think variants.
SAP hoje: +30% de produtividade em P&D, contratação desacelerada, zero headcount novo em 12 meses. Os tokens são cobrados das áreas que os consomem. Chargeback virou variável de RH, e o próprio CEO admite que o gasto não espelha resultado. #FinOps
https://t.co/z4g09ATCeU
Google Cloud knocked it out of the park again. They're now at a ~$99B run rate growing 82% YoY. Another MASSIVE jump in YoY growth
Quarterly YoY growth trends below $GOOG
85% dizem usar IA. 29% incorporaram ao trabalho.
A Atlassian atribui a diferença à titularidade: tirou o
programa de dentro da TI.
A mesma pergunta vale para o custo. De quem é o dashboard de IA na sua empresa?
#FinOps#GovernançaDeIA
https://t.co/jv0x2YD25V
The American Civil Liberties Union of Massachusetts says it’s releasing an online toolkit for attorneys to uncover the technologies police use—and conceal—to build criminal cases, from facial recognition to AI-written police reports. https://t.co/KNKwMdNRws
O NIST Cybersecurity Framework tem 55 páginas. A palavra "backup" aparece em uma delas.
Perguntado sobre backup, um sistema de RAG ranqueou essa página em 55º. Último.
O modelo respondeu com o controle errado. E citou a fonte. Código NIST válido, controle errado.
A gente aprendeu a desconfiar da resposta sem fonte. Ninguém ensinou a desconfiar da resposta com fonte. #RAG #GovernançaDeIA
https://t.co/CxV3dBQonR
Anthropic's Safety Superpower https://t.co/iwiIUTUVT8 via @stratechery What really caught my eye in this @stratechery piece is Anthropic shifting to a 30-day data retention policy—even for enterprise plans that used to promise zero retention. All wrapped up in the name of "safety". Pretty wild move for the enterprise market!
#ArtificialIntelligence #Claude #DataPrivacy #EnterpriseTech #AI
Modelos chineses de IA avançam mais rápido do que o mercado brasileiro monitora. Um economista analisa os grandes players de IA neste vídeo — vale o play. https://t.co/kz7MF1QlUK
#InteligenciaArtificial#China#GovernançaIA
A conta da IA chegou! O Financial Times mostrou que até gigantes como Uber e Amazon estouraram o budget com agentes de IA. O erro não é da tecnologia, é da falta de governança.
Vídeo essencial sobre como controlar custos (FinOps) sem frear a inovação: https://t.co/nnCl9aBrtK via @YouTube
#InteligenciaArtificial #GovernancadeIA #FinOps #Lideranca #Tech
"There’s no asset so good that it can’t become overpriced and thus dangerous, and there are few assets so bad that they can’t get cheap enough to be a bargain" Howard Marks #OvervaluedStocks
"Practice and research show that obedience to a methodology in a cognitive environment only serves to ensure formal coverage for blame, not the success of working results. Methodologies depend on high degrees of predictability to have a high yield. Complex product development does not have that high degree of predictability. It is even more unpredictable than it is predictable." Gunther Verheyen #Agile #Scrum
"You can make more friends in two months by becoming interested in other people than you can in two years by trying to get other people interested in you." Dale Carnegie
🚀 Published this week a new article on LinkedIn about the crucial role of trust in long-term client relationships! Trust me, it's a must-read! 💼🤝 #CustomerSuccess#ClientRelationships#TrustInBusiness#CSM https://t.co/vaP3H96Prj