Heard it more than once: clear pitch that stood out. That took the pressure off immensely.
Next stop: Slush 2025, Helsinki.
Heard it more than once: clear pitch that stood out. That took the pressure off immensely.
Next stop: Slush 2025, Helsinki.
Late invoices don't need to take this much time and money. Now they won't.
Tomorrow: Tirana Slush'd. Then Slush Helsinki.
Big thanks to the teams and everyone who believes in us.
Late invoices don't need to take this much time and money. Now they won't.
Tomorrow: Tirana Slush'd. Then Slush Helsinki.
Big thanks to the teams and everyone who believes in us.
"My business is making money. But I'm running out of cash."
Turns out? They're not alone. 22% of UK businesses are in exactly that position right now.
"My business is making money. But I'm running out of cash."
Turns out? They're not alone. 22% of UK businesses are in exactly that position right now.
Last week, OpenAI launched GPT-5 Codex, built specifically to challenge GitHub Copilot and Cursor AI. This isn't just another coding assistant - it's a direct attack on Microsoft's developer tools empire.
Last week, OpenAI launched GPT-5 Codex, built specifically to challenge GitHub Copilot and Cursor AI. This isn't just another coding assistant - it's a direct attack on Microsoft's developer tools empire.
Stanford researchers put hard numbers on what managers suspected: employment for 22-25 year olds in AI-exposed roles dropped 13% since 2022. Older workers in those same jobs grew 6-12%.
Stanford researchers put hard numbers on what managers suspected: employment for 22-25 year olds in AI-exposed roles dropped 13% since 2022. Older workers in those same jobs grew 6-12%.
OpenAI published research explaining why AI confidently states false information even when it has access to correct data.
The answer is simple: AI gets rewarded for guessing instead of saying "I don't know."
OpenAI published research explaining why AI confidently states false information even when it has access to correct data.
The answer is simple: AI gets rewarded for guessing instead of saying "I don't know."
Obligations for providers of general-purpose AI models have already kicked in as of August 2, 2025, moving talk of transparency and oversight into enforceable requirements.
Obligations for providers of general-purpose AI models have already kicked in as of August 2, 2025, moving talk of transparency and oversight into enforceable requirements.
Today’s signal is hard to miss: utilities are staring at nearly 400 gigawatts of new data center power requests, more than half of America’s peak demand, much of it tied to AI. Many of these sites may never be built, but there is a queue.
Today’s signal is hard to miss: utilities are staring at nearly 400 gigawatts of new data center power requests, more than half of America’s peak demand, much of it tied to AI. Many of these sites may never be built, but there is a queue.
Two days ago, he told reporters "investors as a whole are overexcited about AI."
Tech stocks immediately crashed. Nvidia dropped 3.5%. Palantir fell 10%.
But isn't he the main reason behind this overexcitement?
Two days ago, he told reporters "investors as a whole are overexcited about AI."
Tech stocks immediately crashed. Nvidia dropped 3.5%. Palantir fell 10%.
But isn't he the main reason behind this overexcitement?
Last Wednesday, Google announced a three-year plan to provide AI training and tools to over 100 US colleges and universities.
This is not just corporate philanthropy. This is strategic workforce development.
Last Wednesday, Google announced a three-year plan to provide AI training and tools to over 100 US colleges and universities.
This is not just corporate philanthropy. This is strategic workforce development.
I spent days testing it against Claude Code and o3 on complex data problems.
Without thinking mode, ChatGPT-5 performed worse than both competitors.
With thinking mode enabled, it suddenly became the top performer.
1/3
I spent days testing it against Claude Code and o3 on complex data problems.
Without thinking mode, ChatGPT-5 performed worse than both competitors.
With thinking mode enabled, it suddenly became the top performer.
1/3
The specs: 120b runs on a high-end computer with a good GPU. 20b needs just 16GB memory and can run on a phone. Both use mixture-of-experts so they activate fewer parameters than their total size suggests.
The specs: 120b runs on a high-end computer with a good GPU. 20b needs just 16GB memory and can run on a phone. Both use mixture-of-experts so they activate fewer parameters than their total size suggests.
After six years of closed models, OpenAI just dropped gpt-oss-120b and gpt-oss-20b under Apache 2.0 license.
After six years of closed models, OpenAI just dropped gpt-oss-120b and gpt-oss-20b under Apache 2.0 license.
Remote workers spend 4.5 hours daily in focused work. Office workers manage 3.7 hours. That extra focus time adds up to 4+ hours of productive work per week.
Remote workers spend 4.5 hours daily in focused work. Office workers manage 3.7 hours. That extra focus time adds up to 4+ hours of productive work per week.
The numbers from 50,000 workers tell an interesting story.
Average workday dropped by 36 minutes compared to pre-remote work. Yet productivity increased by 2%. Workers now end their day at 4:39 PM instead of 5:21 PM.
The numbers from 50,000 workers tell an interesting story.
Average workday dropped by 36 minutes compared to pre-remote work. Yet productivity increased by 2%. Workers now end their day at 4:39 PM instead of 5:21 PM.
Then DeepMind said they got gold too.
Except neither actually participated in the official competition.
Then DeepMind said they got gold too.
Except neither actually participated in the official competition.
The Windsurf story ended last week when Cognition Labs bought what was left.
The Windsurf story ended last week when Cognition Labs bought what was left.
Once you've seen bread double in price every week, you stop trusting price tags fast.
I feel that same shock whenever ChatGPT spits out an answer. Ideas that took years to learn now arrive for free, in seconds.
Once you've seen bread double in price every week, you stop trusting price tags fast.
I feel that same shock whenever ChatGPT spits out an answer. Ideas that took years to learn now arrive for free, in seconds.
We’re moving from the Blackberry era to the final form factor, and one of the biggest releases is rumoured in July.
We’re moving from the Blackberry era to the final form factor, and one of the biggest releases is rumoured in July.