Computers can now program themselves, says Jensen Huang.
He says you just have to show a computer what you want, but digital technology cannot replace physical technology. I like this framing: AI can start from the outcome we want, while the real world still has its own demands.
the desired result clearly and include physical constraints in the task.
He says you just have to show a computer what you want, but digital technology cannot replace physical technology. I like this framing: AI can start from the outcome we want, while the real world still has its own demands.
the desired result clearly and include physical constraints in the task.
0
0
0
15
80% chance of a good AI outcome, estimates Elon Musk, with a sharp caveat.
He says AI is more likely to be “super awesome” or “super bad” than somewhere in the middle. That forecast is optimistic, but not a reason to treat AI’s trajectory as settled.
Builders should pursue le making safety and oversight part of the work.
He says AI is more likely to be “super awesome” or “super bad” than somewhere in the middle. That forecast is optimistic, but not a reason to treat AI’s trajectory as settled.
Builders should pursue le making safety and oversight part of the work.
2
0
5
183
Digital superintelligence could raise Earth’s economic output by about 30%.
Elon Musk gives that estimate, then says robots and lunar mass drivers could potentially expand the economy to a million times Earth’s size.
The useful distinction is between a productivity estimate and vision beyond Earth. Builders can test AI’s impact in one workflow and keep civilization-scale forecasts separate from measured results.
Elon Musk gives that estimate, then says robots and lunar mass drivers could potentially expand the economy to a million times Earth’s size.
The useful distinction is between a productivity estimate and vision beyond Earth. Builders can test AI’s impact in one workflow and keep civilization-scale forecasts separate from measured results.
0
0
6
133
“Complete reset,” says Jensen Huang: he argues countries are at the same starting point in AI.
He urges Germany to seize the moment and leapfrog in computing.
The encouraging part is that past distance from the frontier need not decide who benefits, though an open opportunity Builders can assess one workflow where AI might create value now, without assuming adoption alone creates an advantage.
He urges Germany to seize the moment and leapfrog in computing.
The encouraging part is that past distance from the frontier need not decide who benefits, though an open opportunity Builders can assess one workflow where AI might create value now, without assuming adoption alone creates an advantage.
5
0
7
273
Demis Hassabis puts AGI by 2030 at roughly 50/50, not a promised arrival date.
He asks whether AI can keep up with people across ten mental abilities.
A probability and a capability checklist make a sweeping forecast more useful: builders can track progress instead of taking a ork, track specific capabilities and gaps rather than treating an AGI timeline as a delivery commitment.
He asks whether AI can keep up with people across ten mental abilities.
A probability and a capability checklist make a sweeping forecast more useful: builders can track progress instead of taking a ork, track specific capabilities and gaps rather than treating an AGI timeline as a delivery commitment.
3
0
9
231
The durable part of an AI second brain is the files, not the agent.
I’d keep notes and instructions in a repo I own, so I can switch tools without rebuilding everything.
I’d keep notes and instructions in a repo I own, so I can switch tools without rebuilding everything.
Claude, you're fired.
I just moved my fiancée off Claude Code and onto @bot. We canceled her subscription.
Since February, I've told everyone to run their second brain on Claude Code. This is the first time I've pointed someone at something else. x.com/i/article/2112…
8
0
20
387
Bloomberg reports ByteDance, Alibaba and Tencent are ordering over three million Nvidia chips, putting billions in AI capacity at stake.
The orders are worth over $35 billion, and Nvidia says demand far exceeds production plans.
I find that a striking signal of how strongly g AI capacity, and a reminder to plan for access to compute, not just model choice. For your next workload, estimate the capacity you need and identify a fallback if your preferred hardware is unavailable.
The orders are worth over $35 billion, and Nvidia says demand far exceeds production plans.
I find that a striking signal of how strongly g AI capacity, and a reminder to plan for access to compute, not just model choice. For your next workload, estimate the capacity you need and identify a fallback if your preferred hardware is unavailable.
0
0
6
197
DeepSeek V4.1 Flash coded this rainy neon Tokyo street at night in one HTML file. Zero images, zero libraries.
4.5 minutes of generation, $0.05. It needed one fix.
The exact prompt, try it on your model:
"A rainy neon street in Tokyo at night seen from street level: layered wing flickering signs, wet asphalt with neon reflections, rain with splashes, a train passing on an elevated track every few seconds, drifting fog and a slow camera dolly."
4.5 minutes of generation, $0.05. It needed one fix.
The exact prompt, try it on your model:
"A rainy neon street in Tokyo at night seen from street level: layered wing flickering signs, wet asphalt with neon reflections, rain with splashes, a train passing on an elevated track every few seconds, drifting fog and a slow camera dolly."
5
0
14
253
Sundar Pichai just said Gemini 4 Argon did almost 25 trillion tokens in its first 72 hours.
That's about 3x what Gemini 3.0 did in the same window. And Google is capacity constrained serving it.
Everyone is short on compute right now, not just Google. If you're building on mits before you plan for scale.
That's about 3x what Gemini 3.0 did in the same window. And Google is capacity constrained serving it.
Everyone is short on compute right now, not just Google. If you're building on mits before you plan for scale.
3
0
10
578
Set a realistic speed target by measuring the slowest part of the path.
In the clip, a built-in NVMe test reports 7 GB/s. The Linux image is 5.8 GB, but the USB key is slower than the drive, so it sets the theoretical installation limit. A roughly 12-second Dell XPS install is a n here.
My interpretation: the useful goal isn’t an arbitrary speedup; it’s reducing overhead until the workflow approaches its real constraints. For your next slow transfer or install, measure the source and destination paths, find the slower one, and optimize the steps between them.
That gives you a grounded target instead of blind tuning.
Clip: youtube.com/watch?v=NYFGCE…
In the clip, a built-in NVMe test reports 7 GB/s. The Linux image is 5.8 GB, but the USB key is slower than the drive, so it sets the theoretical installation limit. A roughly 12-second Dell XPS install is a n here.
My interpretation: the useful goal isn’t an arbitrary speedup; it’s reducing overhead until the workflow approaches its real constraints. For your next slow transfer or install, measure the source and destination paths, find the slower one, and optimize the steps between them.
That gives you a grounded target instead of blind tuning.
Clip: youtube.com/watch?v=NYFGCE…
2
0
7
240
Turn proposal corrections into a reusable starting point for your next draft.
The clip shows the problem: a new agent session may not remember your preferences. After refining a proposal, ask Claude to save the process as a skill, including your corrections and the finished PDF youtube.com/watch?v=5p-sq8…)
I see this as turning feedback into process knowledge, not just polishing one document. Reuse the skill as a guide for the next proposal, then review the draft: your corrections become a starting point, not a substitute for judgment.
The clip shows the problem: a new agent session may not remember your preferences. After refining a proposal, ask Claude to save the process as a skill, including your corrections and the finished PDF youtube.com/watch?v=5p-sq8…)
I see this as turning feedback into process knowledge, not just polishing one document. Reuse the skill as a guide for the next proposal, then review the draft: your corrections become a starting point, not a substitute for judgment.
3
0
7
271
Use a cloud agent for the conversation and a local model for recurring checks.
In this clip, the creator proposes local models for slower-tolerant work like checking logs or emails, then shows a cloud model orchestrating agents while subagent runs continue in the background.
My s. cloud is a task-allocation choice, not an all-or-nothing bet. Keep the responsive cloud model in the interactive path; try a local subagent for a recurring check whose result can wait, then review what it finds.
In this clip, the creator proposes local models for slower-tolerant work like checking logs or emails, then shows a cloud model orchestrating agents while subagent runs continue in the background.
My s. cloud is a task-allocation choice, not an all-or-nothing bet. Keep the responsive cloud model in the interactive path; try a local subagent for a recurring check whose result can wait, then review what it finds.
7
1
13
356
Turn one recurring decision into a timely recommendation grounded in your calendar and preferences.
In this clip, an agent checks whether a trail run is scheduled, then uses local weather, saved trails, and available time to recommend an option and message it. The workflow h a daily schedule trigger.
The useful shift, in my view, is giving AI a recurring decision plus the context that makes its answer relevant. For a first workflow, pick one low-stakes decision, connect only the needed context, set its trigger, and have it send a recommendation for you to act on.
In this clip, an agent checks whether a trail run is scheduled, then uses local weather, saved trails, and available time to recommend an option and message it. The workflow h a daily schedule trigger.
The useful shift, in my view, is giving AI a recurring decision plus the context that makes its answer relevant. For a first workflow, pick one low-stakes decision, connect only the needed context, set its trigger, and have it send a recommendation for you to act on.
2
0
7
282
Compare agent runs on answer quality, latency, and token use with this practical evaluation pattern.
The clip builds an inventory-agent eval from questions paired with expected answers, using semantic comparison instead of exact string matching.
It publishes the metric to a ere the builder can inspect latency and total token usage, then compare GPT-OSS and Qwen runs.
I see the key shift as replacing memorable one-off outputs with a repeatable way to judge changes against the task.
For an agent you plan to tune or switch models on, collect representative questions and expected answers, then compare quality, latency, and token use before choosing a configuration.
Watch the short clip: youtube.com/watch?v=D74el9…
The clip builds an inventory-agent eval from questions paired with expected answers, using semantic comparison instead of exact string matching.
It publishes the metric to a ere the builder can inspect latency and total token usage, then compare GPT-OSS and Qwen runs.
I see the key shift as replacing memorable one-off outputs with a repeatable way to judge changes against the task.
For an agent you plan to tune or switch models on, collect representative questions and expected answers, then compare quality, latency, and token use before choosing a configuration.
Watch the short clip: youtube.com/watch?v=D74el9…
2
1
8
254
Turn an incident runbook into a reviewable investigation: filter logs, split checks, and bring the findings into one report.
The clip shows an agent using connected observability and code-change tools with a runbook skill, filtering logs in code, and delegating telemetry and sub-agents before a lead agent combines the findings.
My interpretation: the leverage is a repeatable investigation shape, not AI for its own sake. For your first workflow, connect the relevant tools, encode the steps in a runbook, and ask for likely cause, supporting evidence, and next steps for the on-call engineer to review.
The demonstrated output is a shareable file.
The clip shows an agent using connected observability and code-change tools with a runbook skill, filtering logs in code, and delegating telemetry and sub-agents before a lead agent combines the findings.
My interpretation: the leverage is a repeatable investigation shape, not AI for its own sake. For your first workflow, connect the relevant tools, encode the steps in a runbook, and ask for likely cause, supporting evidence, and next steps for the on-call engineer to review.
The demonstrated output is a shareable file.
2
0
7
325
Prototype a voice agent visually, then move into generated Python when you need custom logic.
The clip shows Agent Builder configuring providers and prompts, generating downloadable Python, and letting you configure tools and test the agent in the interface. Then open the code ools, workflows, and custom logic.
I see the leverage in using the interface to shape the first version without making it the limit of the workflow. Start there to configure and test; when the agent needs custom behavior, take the generated Python into your editor.
[Clip](youtube.com/watch?v=n-hCOs…)
The clip shows Agent Builder configuring providers and prompts, generating downloadable Python, and letting you configure tools and test the agent in the interface. Then open the code ools, workflows, and custom logic.
I see the leverage in using the interface to shape the first version without making it the limit of the workflow. Start there to configure and test; when the agent needs custom behavior, take the generated Python into your editor.
[Clip](youtube.com/watch?v=n-hCOs…)
5
0
9
267
Spend 15-20 minutes building a coding-agent brief before handing off a feature.
The clip shows Claude getting a codebase tour, requirements, specifications, and constraints. In a planning conversation, it explores files and helps identify the desired outcome, files to change,
My interpretation: the leverage is moving discovery and alignment ahead of implementation. Before you delegate, write down the outcome, likely files, patterns, requirements, and constraints. That gives the agent a bounded task instead of a vague request to guess at.
Source: youtube.com/watch?v=fHWFF_…
The clip shows Claude getting a codebase tour, requirements, specifications, and constraints. In a planning conversation, it explores files and helps identify the desired outcome, files to change,
My interpretation: the leverage is moving discovery and alignment ahead of implementation. Before you delegate, write down the outcome, likely files, patterns, requirements, and constraints. That gives the agent a bounded task instead of a vague request to guess at.
Source: youtube.com/watch?v=fHWFF_…
11
1
13
342
Build a basic tool-using agent by designing the handoff, not just the tool.
The clip shows the loop: user input, planning, action, observation, then output. It also makes the developer’s role explicit: provide the observation after the tool acts.
My interpretation: a tool call he observation handoff is what lets the model use what happened. When building an agent, wire that return path into the core flow.
Start with the clip’s simple example: a weather tool that takes a city and returns weather details.
The clip shows the loop: user input, planning, action, observation, then output. It also makes the developer’s role explicit: provide the observation after the tool acts.
My interpretation: a tool call he observation handoff is what lets the model use what happened. When building an agent, wire that return path into the core flow.
Start with the clip’s simple example: a weather tool that takes a city and returns weather details.
5
0
8
218
Same query: 24,000 tokens before context optimization, 2,400 after. (youtube.com/watch?v=tZycGA…)
The speaker calls that almost 90% savings. It’s a demo, not a guarantee for your workflow. (evidence: the same-query runs used 24,000 tokens, then 2,400; the speaker described the My interpretation: context is a cost you can optimize, not just a fixed price of using an agent. For a costly workflow, keep the query fixed, toggle the optimizer, and compare token use before redesigning the task or model.
The speaker calls that almost 90% savings. It’s a demo, not a guarantee for your workflow. (evidence: the same-query runs used 24,000 tokens, then 2,400; the speaker described the My interpretation: context is a cost you can optimize, not just a fixed price of using an agent. For a costly workflow, keep the query fixed, toggle the optimizer, and compare token use before redesigning the task or model.
7
0
15
400