Classic prompt injection exfiltration attack here, reported against Salesforce Agentforce and now fixed
PROMPT ENGINEERING
-
Flexible parsing unlocks hidden capabilities in weaker language models
By
–
Flexible parsing revealed hidden capabilities in weaker models LLM-based parsers captured valid answers from free-form reasoning, improving reported accuracy for models that struggled with rigid constraints. It’s a trade-off: exactness versus resilience.
-
Structured Formats Limit AI Reasoning Capabilities
By
–
Structured formats can actually constrain reasoning Models like GPT-4.1 and Grok-3 performed worse when forced into rigid JSON structures. The formatting requirements limited their ability to think through complex problems.
-
OpenAI Deep Research Tool Transparency and System Prompt Access
By
–
I would use OpenAI Deep Research (and equivalent products from other labs) a whole lot more if I'd seen the full list of tools that are available to them – and ideally their system prompts as well but that's less valuable to me than the tool definitions
-
AI/BI Genie Answers Complex Multi-Step Business Questions
By
–
[Demo] How AI/BI Genie answers complex multi-step questions, understands business terminology, and generates queries for validity review.
— Databricks (@databricks) 26 septembre 2025
Learn how to:
-Search for data and AI assets with Databricks One
-Create visualizations from natural language
-Use Genie for conversational… pic.twitter.com/poBEwjFwuw[Demo] How AI/BI Genie answers complex multi-step questions, understands business terminology, and generates queries for validity review. Learn how to:
-Search for data and AI assets with Databricks One
-Create visualizations from natural language
-Use Genie for conversational -
Taking Extra Care with Model Decoding and Out-of-Distribution Constraints
By
–
okay maybe the question is – how would one take “extra care”? the D is the D, cant be touched. and yet people are constrained decoding thousands of tokens with presumably only slight intelligence drop. maybe my json example is too light of a constraint. the OODness would
-
Poe adds conversation branching for prompt testing
By
–
New: You can now branch conversations from any message on Poe. Each branch copies the conversation up to that point into a new chat so you can test prompts, actions, or ideas and steer the direction without starting over. (1/2)
-
Models Writing Tool Calls in Code Is More Expressive Than JSON
By
–
That's a good direction! Let models write their tool calls in code snippets, it's much more flexible & readable & expressive than the lame JSON blobs of usual tool calling!
-
OpenAI reasoning system answer selection without test-time harness
By
–
OpenAI say "the reasoning system selected which answers to submit with no bespoke test-time harness whatsoever" – but were the answers themselves developed with the assistance of tools?
-
Vibe Coding: Building Complete Apps with Single Prompts
By
–
Vibe coding is evolving fast. With a single prompt, you can:
— Abacus.AI (@abacusai) 26 septembre 2025
•Design a frontend
•Set up user login & permissions
•Build an editor
•Spin up and connect a database
You can’t replace real developers, but you can build a simple, usable app for collaboration, tracking, and… pic.twitter.com/asQcWETD3iVibe coding is evolving fast. With a single prompt, you can:
•Design a frontend
•Set up user login & permissions
•Build an editor
•Spin up and connect a database You can’t replace real developers, but you can build a simple, usable app for collaboration, tracking, and