2015: move fast and break things
2026: move fast and ask the agent if it broke things.
it says no.
it broke things
the new weekly cycle:
monday: burn the whole weekly limit by lunch
tuesday: switch to the small model and pretend it's fine
wednesday: refresh x for a reset announcement
thursday: reset drops, ship 3 weeks of work in 9 hours
friday: limit hit again
Companies used to run like this:
Humans in functions.
Work handed off between teams.
Managers coordinating people.
Now the highest-output teams look different.
This is why the old org chart is becoming less useful.
McKinsey is calling the replacement an “orchestration chart”: assemble the right humans and agents around the outcome, not the department.
The bottleneck is no longer “can the agent do the work.”
It’s “who owns the outcome, and when does a human need to step in?”
AWS, Google, Microsoft and 12 others are giving startups over $1M in free credits right now.
and yes, aws credits work on claude (via bedrock) and azure credits work on gpt.
cloud:
google cloud: up to $350K
cloudflare: up to $350K
aws activate: up to $200K
microsoft for startups: up to $150K
gpus + ai infra:
together ai: up to $50K
elevenlabs: 33M characters free for a year
runpod: $1K
lambda: $1K (nvidia inception members)
dev tools:
datadog: up to $100K
retool: up to $60K
posthog: $50K
vercel: up to $30K
github: $10K
notion: up to 6 months free
stripe atlas: $2.5K in stripe credits
bookmark this. links below 👇
claim here:
google cloud: cloud.google.com/startup
cloudflare: cloudflare.com/startups
aws activate: aws.amazon.com/startups/cred…
microsoft: startups.microsoft.com
together ai: together.ai/startup-accelera…
elevenlabs: elevenlabs.io/startup-grants
runpod: runpod.io/startup-program
lambda: lambda.ai/nvidia-inception
datadog: datadoghq.com/partner/datado…
retool: retool.com/startups
posthog: posthog.com/startups
vercel: vercel.com/startups
github: github.com/enterprise/startu…
notion: notion.com/startups
stripe atlas: stripe.com/atlas
most need you to apply, so read the eligibility
YC, a16z, Techstars and others are funding up to US$2 mil right now.
closing soon:
a16z speedrun: $500K upfront, up to $1M. nov 1
startmate: A$120K. nov 1
y combinator: $500K. nov 2
ERA nyc: $150K. nov 2
techstars: $220K, 8 tracks incl. remote. nov 18
rolling:
pearx: $500K–$2M, still taking late apps
EWOR: €500K for teams with traction
AI House: up to $600K
iterative: $150K–$500K
entrepreneurs first: up to $250K
norrsken evolve: up to €500K
z fellows: $10K, one week
opening soon: south park commons, tinyseed, berkeley skydeck
bookmark this before nov 1. links below 👇
apply here:
a16z speedrun: speedrun.a16z.com/apply
y combinator: ycombinator.com/apply
ERA: eranyc.com/apply
techstars: techstars.com/accelerators
startmate: startmate.com/accelerator
pearx: pear.vc/pearx-application
EWOR: ewor.com/apply
AI House: apply.aihouse.vc
iterative: iterative.vc/apply
entrepreneurs first: apply.joinef.com
norrsken evolve: norrskenevolve.vc
z fellows: zfellows.com
bonus, both rolling:
antler: antler.co/apply
accel atoms: atoms.accel.com
terms change, so double check before you sign
overheard in sf:
'so what does your startup do'
'we're an ai agent for ai agents'
'who pays'
'ai agents'
'do they have money'
'not yet'
Another new feature that we’ll have soon is each ChatGPT Chat mode having its own VM. Rolling out by batches
GPT-6 Astra Pro in ChatGPT got an upgrade with faster responses and improved use of plugins and tools.
And now, multiple agents can tackle different parts of a task in parallel. You can follow each agent’s progress and open completed tasks to review their work.
Now available to Pro users in the Chat tab.







