@thorstenball @bedesqui Specifically check out virtual objects: #virtual-object" target="_blank" rel="nofollow noopener">docs.restate.dev/foundations/se…
English
Greg Kogan
3.5K posts
Greg Kogan
@gregkogan
Eng turned messaging & positioning guy for the most complicated tech. Advocate for “the hard simple.” Grew Pinecone 0 to🦄. Created “vector database” category.
Philly area เข้าร่วม Ağustos 2009
180 กำลังติดตาม1.2K ผู้ติดตาม
What are some products similar to Cloudflare DurableObjects?
Specifically: little bit of compute, long-running network connections, storage, uniquely identifiable?
English
@thogge This is bad. Even worse: putting logos under “customers” heading when they just toyed around with the free plan, or starred the repo, or paused at your conference booth that one time in November…
Old tactics, unfortunately: gkogan.co/dishonest-cust…
English
Misleading “customer” pages (very common) is an easy, reliable tell of a startup that’s either struggling, shady, or both.
English
People, we need to stop doing this. Logos you helped close?! Wtf
English
@nicoalbanese10 Talked to a team just yesterday that’s using restate.dev as part of their custom agent framework. Made by creators of Apache Flink (@StephanEwen) and the streaming folks from Meta.
English
Building durable agents with resumable streams, aborting, and persistence is hard
English
Greg Kogan รีทวีตแล้ว
Introducing Console Inbox: the AI-native service desk that gets better as you use it.
Legacy ticketing systems were built to track work, not get the work done. Tickets pile up, context is missing, and teams spend half their day on triage.
Inbox represents a fundamental shift. Console’s AI resolves every request it can, end-to-end. When it can’t, tickets land in your Inbox with all the context you need to fix the issue fast.
The best part: AI recognizes tickets with recurring patterns and creates new automations so they never become tickets again.
This is the future. IT’s time to build.
Read more here: console.com/blog/inbox-ai-…
English
@startupandrew That’s why Mac Minis are flying off the shelves.
English
Cool, cool...
...but what happens when I close my laptop?
Claude@claudeai
New in Cowork: scheduled tasks. Claude can now complete recurring tasks at specific times automatically: a morning brief, weekly spreadsheet updates, Friday team presentations.
English
@hamburger I’d love to see it. I get nothing but trite try-hard taglines.
English
@SherryYanJiang “Isn’t there yet” doesn’t even begin to convey how not there they are. For the vast majority of Americans outside SF or NY, “ChatGPT” is still “that thing kids use to cheat in school.”
English
the sf/twitter groupthink builds this alternate reality where agent orchestration is everywhere, openclaw this and that.
then ppl act shocked when the rest of the world… isn’t there yet.
English
There's a reason most AI SDR cos pick some vague metric like "pipeline influenced" or "total pipeline added" when they talk about results. This is because they are incentivized on open rates
Most people open these AI spam emails to block the domain, so open rate is high. Pipeline is anything these days, so these cos count open rates towards pipeline
English
@VivaLaPanda @paularambles Both are true. Positioning it as an AI SWE raises expectations, which the product didn’t meet.
English
@paularambles that wasn’t the issue with Devon, it just wasn’t very good
English
Startups waste so much time & energy on site pages without stopping to think if they're needed.
One-product companies don't need a product page. The homepage _is_ the product page. Details can be found in the docs.
Whoever made new @braintrust site gets this. Many don't.
English
I expect agents will continue relying in large part on social proof when choosing tools. Both from trusted humans and trusted agents. Question is how they’ll measure it. What’s the PageRank for agents?
brian flynn@Flynnjamm
English
Today the best example of proxy positioning is @mercor_ai...
Their public materials address experts, but the real message is to AI researchers: "Look, we can convince the world's best experts to label your data."
Greg Kogan@gregkogan
Snyk was once a masterful example of what I call proxy positioning: Messaging to one audience to actually appeal to another. They won devs to better sell to security. Sadly, today they're an example of what I call analyst-brained positioning: Unleash! Fabric! Intelligent layer!
English
Snyk was once a masterful example of what I call proxy positioning: Messaging to one audience to actually appeal to another. They won devs to better sell to security.
Sadly, today they're an example of what I call analyst-brained positioning: Unleash! Fabric! Intelligent layer!
English
@HarryStebbings Hence the appeal of "context graphs" as a moat. Customers can migrate data but not the activity & decision history. ashugarg.substack.com/p/ais-trillion…
English
What no one is talking about that everyone should be talking about is how agents will remove the switching cost of SaaS providers, turning previously irreplaceable systems into carcasses of data repositories.
That is when SaaS will really be hit.
English
@rauchg Yes, and: every 1 hot-name early customer is worth >1,000 no-name customers. Do everything you can to win former early on so you can name drop in your launch. Trust from respected names is the moat.
English
From now on, hype-centric splashy launches will likely be strongly uncorrelated with success.
If by the time you launch you don’t have escape velocity, you will likely get Sybil attacked¹. Agents will spin up 10 competing products with your same interface.
Start with an audience of 1 and get confirmation that it works. Then expand to a circle of friends or design partners. By the time you go public, your moat needs to be deeper than it’s ever been.
¹ en.wikipedia.org/wiki/Sybil_att…
English
Can't believe I'm saying this but, as of today, Codex 5.3 is leagues above Opus 4.6 for messaging & product marketing work.
Claude was great last week but seemingly regressed: now forgets context, gets confused, then falls back to vapid & long-winded marketing drivel. 🤮
English