This is cool! Do you have any plans to add open source capability, maybe through something like LangChain?

We have lots of compute, but no OpenAI api haha

For sure, one of the points of doing Geppetto was to speed up the link between Slack and LLMs or other interfaces.

I am sure that everyone capable can do a similar project but the concern is that developers and organizations want to lower the learning curve.

Digital Ocean recently launched a NVIDIA H100 GPU cluster which provides on-demand (pay-by-the-hour) H100 cloud GPUs backed by a 3.2TB GPU fabric. The cluster is optimal for bursty training/fine-tuning jobs, no commit required.

Server specs (8-way box) 8 NVIDIA H100 HGX GPUs connected by NVLink/NVSwitch 1.76 TB RAM 128 vCPU

DigitalOcean now has H100 GPU clusters available with 3.2TB networking

https://twitter.com/digitalocean/status/1747974629572911140

This might have to be my next DIY project...

Spot on analysis. It's a broken product, most of the time, for me.

The place I work at now uses Teams, and I legitimately didn't notice anything amiss.

- Channels randomly taking several minutes to load or outright refusing to load? Normal

- Calls dropping on a fast, stable ethernet connection? Normal

- Messages not sending, or appearing to send but silently never arriving/being dropped upon arrival? Normal

- Messages double- and triple-sending? Normal

- Messages being sent out-of-order? Normal

- Messages sending extremely slowly? Normal

- Attachments not loading? Normal

- Teams deciding logging in is just too difficult and I have to restart it at least once? Normal

I remember when I got an email today about an apparently outage describing all of the above (minus the call instability), I was like "wait, it's not supposed to be like this?"

And now that I type this out, especially having used Slack in the past, I realize what an indictment of Teams that is. (But I've also worked at a place that used Lync/Skype for Business/whatever they're calling it now, and it still manages to be pleasant compared to that mess. Though I won't give MS any credit there; the bar was on the floor, and they managed to avoid tripping over it.)

And yet companies don't mind the extremely obvious loss of productivity this causes. They also don't mind that it makes them look unprofessional. At all.

I've seen townhall meetings where 200+ FTE where sitting around waiting for Teams to stop glitching for like 10 minutes. And after they go back to their desk for real work, the glitching continues.

"innovation"

My previous employer forced us to use a Hip hat instance. It was legitimately hell, deleting history randomly and just atrocious. Using Teams now (switched jobs) it legitimately feels mystical in comparison.

I can't help but imagine what it'll feel like if you find a job that uses Slack.

I use Slack outside work, and it's less buggy than Teams but the feature count is dramatically lower. Maybe as I use it mostly with non-technical groups. Feels 'fine' but people seem to clamor for Discord.

My job use Slack and I find it quite frustrating. The UI could definitively need some work. Seems like they change the UI now and then but it doesn't really make it easier for me.

Every week I face a new issue with Teams because every week MS will push a silent update which will invariably fix one thing but break another.

The best is when those “silent updates” restart Teams while you are in the middle of using it.