@RachelReevesMP 0.1% is technically growth. But calling this a stronger, more secure economy when construction and production are both contracting feels generous.
We’re rolling out some big improvements to Gemma 4, fueled by incredible community feedback and contributions!
Here is a breakdown of what’s being fixed and updated in this release: 🧵👇
@DWeisburd@APompliano The ego point lands. Playing not to lose looks sensible to everyone watching. That's why so few founders build something that matters.
@itsjulespark@perceptroninc In-context learning without fine-tuning is what stands out. One reference image and it finds every match across new video. An entire annotation pipeline becomes a single API call.
London's IPO pipeline has been functionally closed for growth companies since 2021. Abolishing stamp duty is a start, but founders I speak to are already looking at where capital actually sits now.
"The situation on the London market is now so serious that it requires bolder interventions to save our stock market."
With Andy Burnham set to take office next week, our CEO Richard Stone is calling for action, including abolishing stamp duty on UK shares, boosting investment and reversing cuts to VCT tax relief to help revive UK markets and support future IPOs.
Read more: https://t.co/zLD3PANlcd
#investmenttrusts #investing #investmentcompanies
WILL CAIN'S WARNING: "GET OUT OF BLUE STATES”
@willcain says policies driving high earners and businesses away are creating a dangerous cycle. "The millionaires, the billionaires are leaving... that's how you circle the drain."
@dkundel Had a similar moment last week. Fed it a dashboard screenshot and got working filters and dropdowns back. That jump from static to interactive is where it gets interesting.
Speed on low-cost hardware is what actually matters for deployment. Most manipulation research optimises for success rate and ignores cycle time entirely.
@mihail_eric The pace here tells you everything. A course that launched last September already needs a full rewrite. The gap between people building these tools and people using them is only getting wider.
@oliverbrocato This is what building actually looks like. The hires that don't make 90 days, the money burned before anything sticks. Most people only post the chart.
@afscott@MetaDAOProject@superteam@NiceGuyShri T+0 settlement across emerging markets solves real friction. $350M in five months tells me the demand was already sitting there waiting for proper rails.
@mrp@ttunguz Early stage, the founders I work with aren't tracking compute against salary. They're weighing it against the next hire. Different math entirely.
I keep noticing the same pattern: AI tools optimise the visible layer for demos. Data portability, recovery economics, real SLA coverage - the boring structural stuff decides long-term value.
this seems super interesting! def a lot to learn here
CoreWeave's SLA is a credit-only floor that doesn't cover what actually breaks training. Its 99.9% tier only pays if you're redundant across 2+ regions AND all go down at once (not really training realistic...). So they now sell straggler detection as a feature instead of insuring uptime.
at least they have an SLA though, most neoclouds don't :(
@UKDecline 170,000 people aged 16-24 too sick to work. Before their careers have even started. Mental health support has to reach people earlier, not after they've already fallen out.