@pierceboggan@paleonormie@code Additionally, a centralized way to add Skills, Instructions, etc for Enterprise use, so that we can make those things automatically available for everybody in our org.
The Insights could be a lot better. Cursor has done a good job of developing their analytics, particularly for Enterprise use. I want to see a lot more metrics, like MCP tool usage.
Also, user filtering would be great to see to drill down on specific people and see those metrics as well as how many premium requests they've made for the cycle. Cursor allows you to see every request that comes through at a high level.
@ScottAdamsMeet@LahomaLahiri The invitation for this one was updated earlier today due to the pending snowstorm. Thank you for confirming. It will be on 2/7 at the same time/place.
As bad as the defense was at times, the only game in the final stretch that they really couldnβt hold up in the red zone was against SF. Outside of that game, they largely managed to limit the damage.
I understand Benβs logic there but I think the defense wasnβt bad enough to warrant the level of aggression on 4th down. The Packer game last week was similar. Passing on those points made the game harder than it needed to be.
@Grummz Yes. I am sick of occasionally pulling up some sports related posts and then having my feed flooded with them for the next week.
Most of the time I do not want that type of content but I get bombarded with it so quickly.
I agree with your criticisms here. The reality is that the Grok models have continued to be subpar for agentic coding use cases compared to the other SOTA providers.
I do like Grok Code for simpler tasks due to its speed and price. But I still can't justify using Grok 4 Fast over 4.5 Haiku/Sonnet/Opus right now.
And I've been testing 4.1 Fast in non-agentic tasks this past week as well but have unfortunately seen that its accuracy and speed have been really bad, even compared to 4 Fast. It is clearly not polished enough yet for them to release more broadly.
I just want a true Grok competitor to the Claude models for agentic coding use cases. I hope we see that by early next year.
@techdevnotes I've been running it through some local use cases via OpenRouter the last couple days and it has honestly been pretty terrible.
It has been much slower than Grok 4 Fast and makes a lot of mistakes.
I think they are refining it yet before a broader release.
@C0unterSignal@StefanMolyneux My wife and I went through this last year. We avoided sleep training at first but our daughter slept so much better and was genuinely happier afterwards. It took less than 2 days.
We followed this book:
Healthy Sleep Habits, Happy Child by Marc Weissbluth M.D.
@skcd42@thdxr For me the larger context is helpful for planning work. I donβt provide nearly as much context when actually having the agent write code.
@StefanMolyneux I met both you and Milo at A Night for Freedom in DC back in 2018. Your speech about Freud was brilliant.
The world has changed so much since that time.