Interesting we have #blinks for technical books.
Prompt Engineering for Generative AI Summary of Key Ideas and Review | James Phoenix, Mike Taylor - Blinkist https://t.co/HMkqsPqBSq
#prompt#engineering
@NotionHQ Why subscribe to notion AI?
I like Notion. Itโs my second brain ๐ง .
But most of the AI tools have access to Notion with connectors.
Why add a redundant subscription when we can have perplexity for multi models or get dedicated Claude, codex or grok which offers much more.
@araseb_@grok Yes. But my opinion is
Birth rate
Economy instability
Every crises
Except recent release of Sakana nothing major.
But Iโm big fan they are keeping up with anime ๐๐๐
Anthropic's Opus 4.7 and 4.8 models are experiencing degraded performance, which is causing a higher rate of failures for users selecting these models in Notion AI.
To mitigate impact, all Anthropic models have been disabled in the model picker and requests have been rerouted to alternative providers. Most users should now be able to continue using Notion AI with minimal disruption, though Anthropic-specific features remain unavailable.
Please refer to https://t.co/2jxfCbG6UZ for the details.
It could be more of a strategic move using Nvidia to evaluate the working of the model compared to TPUs so that they can share the benchmarking and advertise to the client. If you use the model with TPU as an infrastructure, you pay a certain amount and you get this value in return. Whereas if you go with the Nvidia, which is again open for the client to choose, you get this cost and this return on investment. It could be more of a strategic move than running out of compute or GPUs.
@NousResearch not sure if hermas is still a PoC, because the agent has been performing really bad. I found a few bugs. Whenever I run Hermes to keep it running healthy, I have to run Claude Code in parallel on the shell to monitor and fix:
- the gateway issues
- the connectivity issue
- the model configuration issues
- the timeout settings
I have to create multiple cron jobs just to keep it running.
For instance, the below response from Claude code
The /model ds switch will still block for up to 8 seconds for the validation call (that's an upstream Hermes bug that requires a
code fix to make it async).
@WisprFlow Wispr Flow is getting bad to worse. It was the best when I started using it. I was using it like my primary keyboard. Tons of messages across multiple devices on Mac, it is the best. I have been using it more than my keyboard, but on iPhone recently it stopped working or functioning as it should be.
I am not able to do one-click speak. I have to do a lot of switching, and then it still doesn't work. Then I have to open the app and click this button to activate it. I do not understand how to keep it active all the time. If I enable it, it shows in the dynamic island. It's a little annoying there, but I wanted to be always available when I switch it from the keyboard. That should be the trigger point to start Wispr Flow, not to go into the app, enable it, and then go back into the keyboard and select the Wispr Flow option.
Hope the team fixes this, because I have been using it a lot and I want to continue using it. I tried multiple apps. Nothing comes close to the accuracy and the refinement that Wispr Flow does, so please fix this and I'll be your long-term customer.