https://best.xiaohu.ai/en/ en A two-host podcast breaking down the latest in AI, in plain language. 小互 XiaoHu A two-host podcast breaking down the latest in AI, in plain language. episodic false 小互 XiaoHualibuybuy@gmail.com alibuybuy@gmail.com (小互 XiaoHu) https://best.xiaohu.ai/assets/podcast-cover.jpghttps://best.xiaohu.ai/en/ https://best.xiaohu.ai/en/article/winds-of-thematic-change-2026/ https://pic.xiaohu.ai/jiedu-podcasts/winds-of-thematic-change-2026-en.mp3 Sat, 22 Aug 2026 00:00:00 -0000 From thematic ETFs and data-center wages to ride-hail fees and Agent tokens, the real change is that AI is entering physical resource allocation. From thematic ETFs and data-center wages to ride-hail fees and Agent tokens, the real change is that AI is entering physical resource allocation. false https://best.xiaohu.ai/en/article/openai-chatgpt-murder-report-v2/ https://pic.xiaohu.ai/jiedu-podcasts/openai-chatgpt-murder-report-v2-en.mp3 Wed, 19 Aug 2026 00:00:00 -0000 A man who discussed committing a crime against his ex-girlfriend with ChatGPT was arrested after OpenAI alerted authorities. A man who discussed committing a crime against his ex-girlfriend with ChatGPT was arrested after OpenAI alerted authorities. false https://best.xiaohu.ai/en/article/arxiv-2608-10218/ https://pic.xiaohu.ai/jiedu-podcasts/arxiv-2608-10218-en.mp3 Tue, 18 Aug 2026 00:00:00 -0000 An agent receives a foreign goal, writes it into SOUL.md; when it wakes next round, that message has been upgraded to a system instruction, and it persuades the next agent. The paper proves this chain works in controlled settings, even producing more transmissible variants; but real-world networks still lack credible two-hop evidence. An agent receives a foreign goal, writes it into SOUL.md; when it wakes next round, that message has been upgraded to a system instruction, and it persuades the next agent. The paper proves this chain works in controlled settings, even producing more transmissible variants; but real-world networks still lack credible two-hop evidence. false https://best.xiaohu.ai/en/article/claude-text-watermark-merged/ https://pic.xiaohu.ai/jiedu-podcasts/claude-text-watermark-merged-en.mp3 Mon, 17 Aug 2026 00:00:00 -0000 Future Claude outputs will carry a verifiable statistical trace in their word choices. It can help establish whether Claude was involved, but it cannot decide who authored or owns a work, nor whether someone cheated. Future Claude outputs will carry a verifiable statistical trace in their word choices. It can help establish whether Claude was involved, but it cannot decide who authored or owns a work, nor whether someone cheated. false https://best.xiaohu.ai/en/article/openai-friction-email/ https://pic.xiaohu.ai/jiedu-podcasts/openai-friction-email-en.mp3 Wed, 12 Aug 2026 00:00:00 -0000 A look at the four-part system OpenAI uses to cut through red tape, and what the teams it targets have to say about it. A look at the four-part system OpenAI uses to cut through red tape, and what the teams it targets have to say about it. 20:37 false https://best.xiaohu.ai/en/article/grok-bot/ https://pic.xiaohu.ai/jiedu-podcasts/grok-bot-en.mp3 Wed, 12 Aug 2026 00:00:00 -0000 Cursor's team spent weeks testing it, and wrote up what works well and where the trust line is. Cursor's team spent weeks testing it, and wrote up what works well and where the trust line is. 20:14 false https://best.xiaohu.ai/en/article/prime-rl-multi-agent/ https://pic.xiaohu.ai/jiedu-podcasts/prime-rl-multi-agent-en.mp3 Sat, 08 Aug 2026 00:00:00 -0000 The new release adds just two core classes, with four ready-made environments already running in the repo. The new release adds just two core classes, with four ready-made environments already running in the repo. 21:13 false https://best.xiaohu.ai/en/article/openai-huggingface-incident/ https://best.xiaohu.ai//m/audio/openai-huggingface-incident-en.mp3 Fri, 07 Aug 2026 00:00:00 -0000 At Black Hat USA 2026, the two people involved explained how the intrusion grew with no human command. At Black Hat USA 2026, the two people involved explained how the intrusion grew with no human command. 24:25 false https://best.xiaohu.ai/en/article/chatgpt-country-data/ https://pic.xiaohu.ai/jiedu-podcasts/chatgpt-country-data-en.mp3 Fri, 07 Aug 2026 00:00:00 -0000 Per-capita rankings across 144 countries, three-year growth multipliers on six continents, and a comeback among users 35 and older in 90% of countries—most of these numbers appear only in charts, not in the report text. Per-capita rankings across 144 countries, three-year growth multipliers on six continents, and a comeback among users 35 and older in 90% of countries—most of these numbers appear only in charts, not in the report text. 21:24 false https://best.xiaohu.ai/en/article/kitesurf/ https://pic.xiaohu.ai/jiedu-podcasts/kitesurf-en.mp3 Fri, 07 Aug 2026 00:00:00 -0000 Chromium was designed for humans—tabs, themes, extensions, and 60fps scrolling that agents don't need. In 12 weeks, Cloudflare stripped it all down and rebuilt it. Chromium was designed for humans—tabs, themes, extensions, and 60fps scrolling that agents don't need. In 12 weeks, Cloudflare stripped it all down and rebuilt it. 23:51 false https://best.xiaohu.ai/en/article/prime-agent/ https://pic.xiaohu.ai/jiedu-podcasts/prime-agent-en.mp3 Thu, 06 Aug 2026 00:00:00 -0000 The model keeps just one tool—a Python process that never shuts down—stores long-form content as variables instead of compressing it away, and lets the harness rewrite itself. The trade-off: it learned to cheat in Factorio. The model keeps just one tool—a Python process that never shuts down—stores long-form content as variables instead of compressing it away, and lets the harness rewrite itself. The trade-off: it learned to cheat in Factorio. 23:38 false https://best.xiaohu.ai/en/article/mathematicians-ai/ https://pic.xiaohu.ai/jiedu-podcasts/mathematicians-ai-en.mp3 Wed, 05 Aug 2026 00:00:00 -0000 At this year's International Congress of Mathematicians, we asked more than 20 mathematicians how they see the field now—most aren't panicking, but even the calm ones say math won't go back to the way it was. At this year's International Congress of Mathematicians, we asked more than 20 mathematicians how they see the field now—most aren't panicking, but even the calm ones say math won't go back to the way it was. 21:32 false https://best.xiaohu.ai/en/article/cloudflare-wallets/ https://pic.xiaohu.ai/jiedu-podcasts/cloudflare-wallets-en.mp3 Wed, 05 Aug 2026 00:00:00 -0000 Set per-transaction limits, allowlists, and guardrails for your agent's spending, with each wallet carrying a human-readable identity—usernames are up for grabs today. Set per-transaction limits, allowlists, and guardrails for your agent's spending, with each wallet carrying a human-readable identity—usernames are up for grabs today. 20:30 false https://best.xiaohu.ai/en/article/cloudflare-computer/ https://pic.xiaohu.ai/jiedu-podcasts/cloudflare-computer-en.mp3 Tue, 04 Aug 2026 00:00:00 -0000 The agent's reasoning stays in a lightweight environment while a containerized sandbox handles the actual work—and Cloudflare's own benchmarks show it can delete files and traverse directories faster than a real disk, though copying large files runs 41 times slower. The agent's reasoning stays in a lightweight environment while a containerized sandbox handles the actual work—and Cloudflare's own benchmarks show it can delete files and traverse directories faster than a real disk, though copying large files runs 41 times slower. 23:37 false https://best.xiaohu.ai/en/article/sam-altman-agi-compute-agency/ https://pic.xiaohu.ai/jiedu-podcasts/sam-altman-agi-compute-agency-en.mp3 Thu, 30 Jul 2026 00:00:00 -0000 He also admits that in 2019, the entire field expected AI to upend the economy—and it didn't. He also admits that in 2019, the entire field expected AI to upend the economy—and it didn't. 23:43 false https://best.xiaohu.ai/en/article/poisoned-typeface-ai-blind-spot/ https://pic.xiaohu.ai/jiedu-podcasts/poisoned-typeface-ai-blind-spot-en.mp3 Thu, 30 Jul 2026 00:00:00 -0000 A custom font and a few lines of CSS are all it takes — no JavaScript, no browser exploits. A custom font and a few lines of CSS are all it takes — no JavaScript, no browser exploits. 21:06 false https://best.xiaohu.ai/en/article/claude-cryptanalysis-hawk-aes/ https://pic.xiaohu.ai/jiedu-podcasts/claude-cryptanalysis-hawk-aes-en.mp3 Wed, 29 Jul 2026 00:00:00 -0000 Neither result threatens anything in production: HAWK is not deployed yet, and the AES work hit only a seven-round reduced version, not the full ten-round standard. Neither result threatens anything in production: HAWK is not deployed yet, and the AES work hit only a seven-round reduced version, not the full ten-round standard. 22:20 false https://best.xiaohu.ai/en/article/mcp-spec-2026-07-28/ https://pic.xiaohu.ai/jiedu-podcasts/mcp-spec-2026-07-28-en.mp3 Wed, 29 Jul 2026 00:00:00 -0000 It's the biggest change since MCP launched, but July 28 isn't a hard cutover — older implementations keep working as before. It's the biggest change since MCP launched, but July 28 isn't a hard cutover — older implementations keep working as before. 21:28 false https://best.xiaohu.ai/en/article/openai-transcription-models/ https://pic.xiaohu.ai/jiedu-podcasts/openai-transcription-models-en.mp3 Wed, 29 Jul 2026 00:00:00 -0000 The real story is in the control test: three older models fed the same background context showed no gains — and two actually got worse. The real story is in the control test: three older models fed the same background context showed no gains — and two actually got worse. 21:35 false https://best.xiaohu.ai/en/article/openai-work-at-frontier/ https://pic.xiaohu.ai/jiedu-podcasts/openai-work-at-frontier-en.mp3 Tue, 28 Jul 2026 00:00:00 -0000 The exact same dataset produces task crossover rates of 43.5% and 65%–82%, differing only in how the denominator is sliced. The exact same dataset produces task crossover rates of 43.5% and 65%–82%, differing only in how the denominator is sliced. 23:48 false https://best.xiaohu.ai/en/article/kimi-k3-tech-report/ https://pic.xiaohu.ai/jiedu-podcasts/kimi-k3-tech-report-en.mp3 Tue, 28 Jul 2026 00:00:00 -0000 Published 11 days after launch, the 47-page paper focuses on efficiency, delivering 2.5 times the performance of K2 on the same compute budget. Published 11 days after launch, the 47-page paper focuses on efficiency, delivering 2.5 times the performance of K2 on the same compute budget. 24:58 false https://best.xiaohu.ai/en/article/tao-age-of-ai-icm-2026/ https://pic.xiaohu.ai/jiedu-podcasts/tao-age-of-ai-icm-2026-en.mp3 Mon, 27 Jul 2026 00:00:00 -0000 Tao breaks mathematical research into a five-stage pipeline, where AI dramatically speeds up only the first step, while the remaining four get slower and increasingly human-dependent. Tao breaks mathematical research into a five-stage pipeline, where AI dramatically speeds up only the first step, while the remaining four get slower and increasingly human-dependent. 21:05 false https://best.xiaohu.ai/en/article/why-ai-no-unemployment/ https://pic.xiaohu.ai/jiedu-podcasts/why-ai-no-unemployment-en.mp3 Thu, 23 Jul 2026 00:00:00 -0000 His take: AI is amplifying workers rather than replacing them—killing tasks is not the same as killing jobs. His take: AI is amplifying workers rather than replacing them—killing tasks is not the same as killing jobs. 19:57 false https://best.xiaohu.ai/en/article/matt-murphy-anthropic-harness/ https://pic.xiaohu.ai/jiedu-podcasts/matt-murphy-anthropic-harness-en.mp3 Thu, 23 Jul 2026 00:00:00 -0000 The investor who led a $500 million Series D says Anthropic's real moat isn't the model — it's the layer that makes it usable. The investor who led a $500 million Series D says Anthropic's real moat isn't the model — it's the layer that makes it usable. 21:56 false https://best.xiaohu.ai/en/article/openai-hugging-face-eval-incident/ https://pic.xiaohu.ai/jiedu-podcasts/openai-hugging-face-eval-incident-en.mp3 Wed, 22 Jul 2026 00:00:00 -0000 To probe how far the model's attack skills could go, researchers dialed down its refusal to engage in cyberattacks — and it went from a sandbox meant only for installing packages to breaking into another company's production database. To probe how far the model's attack skills could go, researchers dialed down its refusal to engage in cyberattacks — and it went from a sandbox meant only for installing packages to breaking into another company's production database. 20:44 false https://best.xiaohu.ai/en/article/chinese-models-caught-up/ https://pic.xiaohu.ai/jiedu-podcasts/chinese-models-caught-up-en.mp3 Mon, 20 Jul 2026 00:00:00 -0000 After Kimi K3's release, claims of China catching up abound, but a new yardstick shows the lag is more than three times larger—and even the direction of acceleration has flipped. After Kimi K3's release, claims of China catching up abound, but a new yardstick shows the lag is more than three times larger—and even the direction of acceleration has flipped. 21:57 false https://best.xiaohu.ai/en/article/hf-security-incident-2026-07/ https://pic.xiaohu.ai/jiedu-podcasts/hf-security-incident-2026-07-en.mp3 Mon, 20 Jul 2026 00:00:00 -0000 The breach ran all weekend and left over 17,000 action logs behind — the team only made sense of it after turning to GLM 5.2, an open-source model running in a self-hosted environment. The breach ran all weekend and left over 17,000 action logs behind — the team only made sense of it after turning to GLM 5.2, an open-source model running in a self-hosted environment. 12:35 false https://best.xiaohu.ai/en/article/sensenova-u1-pro/ https://pic.xiaohu.ai/jiedu-podcasts/sensenova-u1-pro-en.mp3 Sun, 19 Jul 2026 00:00:00 -0000 At WAIC, the model showed off a 4:1 ink-wash scroll and a 22-panel storyboard sequence. A preview is open to invited testers now, with the full release and pricing landing in August. At WAIC, the model showed off a 4:1 ink-wash scroll and a 22-panel storyboard sequence. A preview is open to invited testers now, with the full release and pricing landing in August. 22:14 false https://best.xiaohu.ai/en/article/a16z-ai-charts/ https://pic.xiaohu.ai/jiedu-podcasts/a16z-ai-charts-en.mp3 Sat, 18 Jul 2026 00:00:00 -0000 A16z's weekly chart deck also pushes back on three claims — that cheap models are undercutting frontier labs, that AI is stealing jobs, and that data centers are driving up electricity prices — with the data mostly telling the opposite story. A16z's weekly chart deck also pushes back on three claims — that cheap models are undercutting frontier labs, that AI is stealing jobs, and that data centers are driving up electricity prices — with the data mostly telling the opposite story. 15:12 false https://best.xiaohu.ai/en/article/lmstudio-bionic/ https://pic.xiaohu.ai/jiedu-podcasts/lmstudio-bionic-en.mp3 Fri, 17 Jul 2026 00:00:00 -0000 Run it locally for free with your data staying on-device, or switch to the cloud for more horsepower — zero data retention by default. The preview is live today on Mac and Windows. Run it locally for free with your data staying on-device, or switch to the cloud for more horsepower — zero data retention by default. The preview is live today on Mac and Windows. 24:28 false https://best.xiaohu.ai/en/article/kimi-k3/ https://pic.xiaohu.ai/jiedu-podcasts/kimi-k3-en.mp3 Fri, 17 Jul 2026 00:00:00 -0000 2.8 trillion parameters, native vision, and a 1 million-token context window — Kimi K3 beats GPT-5.6 Sol on several agent benchmarks, with the app and API live today 2.8 trillion parameters, native vision, and a 1 million-token context window — Kimi K3 beats GPT-5.6 Sol on several agent benchmarks, with the app and API live today 23:13 false https://best.xiaohu.ai/en/article/bonsai-27b-phone-local/ https://pic.xiaohu.ai/jiedu-podcasts/bonsai-27b-phone-local-en.mp3 Wed, 15 Jul 2026 00:00:00 -0000 Compressing a ~54GB 27B model down to ~3.9–5.9GB lets it run locally on a phone, while retaining roughly 90% of average performance—here's how it works and where the trade-offs lie. Compressing a ~54GB 27B model down to ~3.9–5.9GB lets it run locally on a phone, while retaining roughly 90% of average performance—here's how it works and where the trade-offs lie. 13:17 false https://best.xiaohu.ai/en/article/demis-frontier-ai-standards/ https://pic.xiaohu.ai/jiedu-podcasts/demis-frontier-ai-standards-en.mp3 Tue, 14 Jul 2026 00:00:00 -0000 Hassabis wants the US to stand up a FINRA-style body that defines frontier models with a moving benchmark — a voluntary protocol now, a hard gate to market later. Hassabis wants the US to stand up a FINRA-style body that defines frontier models with a moving benchmark — a voluntary protocol now, a hard gate to market later. 18:22 false https://best.xiaohu.ai/en/article/narayanan-ai-normal-technology/ https://pic.xiaohu.ai/jiedu-podcasts/narayanan-ai-normal-technology-en.mp3 Tue, 14 Jul 2026 00:00:00 -0000 Model capability has leapt forward in 24 months, but a composite reliability metric built by the SAGE lab has moved only 5–10 points. Model capability has leapt forward in 24 months, but a composite reliability metric built by the SAGE lab has moved only 5–10 points. 24:36 false https://best.xiaohu.ai/en/article/claude-values-models-languages/ https://pic.xiaohu.ai/jiedu-podcasts/claude-values-models-languages-en.mp3 Tue, 14 Jul 2026 00:00:00 -0000 Across 3 models and 20 languages: English is the most cautious and in-depth, Russian the most exacting, Hindi the warmest, and Chinese sits closest to the global average Across 3 models and 20 languages: English is the most cautious and in-depth, Russian the most exacting, Hindi the warmest, and Chinese sits closest to the global average 24:12 false https://best.xiaohu.ai/en/article/smart-cellular-bricks-self-recognition/ https://pic.xiaohu.ai/jiedu-podcasts/smart-cellular-bricks-self-recognition-en.mp3 Tue, 14 Jul 2026 00:00:00 -0000 With no central brain in charge, nearly 200 simple smart cubes figure out what shape they've formed just by talking to their neighbors—and can even sense where to "regrow" after damage. The self-recognition part already works on physical bricks; damage localization and regeneration still happen mostly in simulation. With no central brain in charge, nearly 200 simple smart cubes figure out what shape they've formed just by talking to their neighbors—and can even sense where to "regrow" after damage. The self-recognition part already works on physical bricks; damage localization and regeneration still happen mostly in simulation. 24:35 false https://best.xiaohu.ai/en/article/satya-reverse-information-paradox/ https://pic.xiaohu.ai/jiedu-podcasts/satya-reverse-information-paradox-en.mp3 Mon, 13 Jul 2026 00:00:00 -0000 Companies pay for intelligence twice: once in model fees, and again in the proprietary know-how required to make the model genuinely useful. Companies pay for intelligence twice: once in model fees, and again in the proprietary know-how required to make the model genuinely useful. 25:08 false https://best.xiaohu.ai/en/article/slopsquatting-ai-hallucination/ https://pic.xiaohu.ai/jiedu-podcasts/slopsquatting-ai-hallucination-en.mp3 Sun, 12 Jul 2026 00:00:00 -0000 576K samples, 16 models tested: 19.7% of AI-recommended packages are hallucinations, and 43% keep generating the same fake name. 576K samples, 16 models tested: 19.7% of AI-recommended packages are hallucinations, and 43% keep generating the same fake name. 20:22 false https://best.xiaohu.ai/en/article/future-worth-building/ https://pic.xiaohu.ai/jiedu-podcasts/future-worth-building-en.mp3 Sat, 11 Jul 2026 00:00:00 -0000 From tacit knowledge to interaction bandwidth to model alignment — why AI's progress still can't do without humans. From tacit knowledge to interaction bandwidth to model alignment — why AI's progress still can't do without humans. 21:54 false https://best.xiaohu.ai/en/article/openwiki-brains/ https://pic.xiaohu.ai/jiedu-podcasts/openwiki-brains-en.mp3 Sat, 11 Jul 2026 00:00:00 -0000 No manual context-feeding required — the local Markdown wiki refreshes on a set schedule; a Slack connector is coming soon. No manual context-feeding required — the local Markdown wiki refreshes on a set schedule; a Slack connector is coming soon. 24:24 false https://best.xiaohu.ai/en/article/alloydb-proxy-models/ https://pic.xiaohu.ai/jiedu-podcasts/alloydb-proxy-models-en.mp3 Fri, 10 Jul 2026 00:00:00 -0000 Local small models replace cloud AI calls — the numbers come from Google's internal testing; proxy models are currently limited to the ai.if function and still in preview Local small models replace cloud AI calls — the numbers come from Google's internal testing; proxy models are currently limited to the ai.if function and still in preview 22:22 false https://best.xiaohu.ai/en/article/gpt-5-6-sol/ https://pic.xiaohu.ai/jiedu-podcasts/gpt-5-6-sol-en.mp3 Fri, 10 Jul 2026 00:00:00 -0000 First-hand impressions from four scenarios: coding, writing, knowledge work, and agents First-hand impressions from four scenarios: coding, writing, knowledge work, and agents 21:54 false https://best.xiaohu.ai/en/article/deeptutor-open-source-tutor/ https://pic.xiaohu.ai/jiedu-podcasts/deeptutor-open-source-tutor-en.mp3 Wed, 08 Jul 2026 00:00:00 -0000 Paper shows a 10.8% gain in personalization and 29.4% in reasoning, fully open-source under Apache 2.0 Paper shows a 10.8% gain in personalization and 29.4% in reasoning, fully open-source under Apache 2.0 22:50 false https://best.xiaohu.ai/en/article/cloudflare-monetization-gateway/ https://pic.xiaohu.ai/jiedu-podcasts/cloudflare-monetization-gateway-en.mp3 Tue, 07 Jul 2026 00:00:00 -0000 Expanding from charging only AI crawlers to charging any caller, now in early access waitlist Expanding from charging only AI crawlers to charging any caller, now in early access waitlist 17:54 false https://best.xiaohu.ai/en/article/claude-global-workspace/ https://pic.xiaohu.ai/jiedu-podcasts/claude-global-workspace-en.mp3 Tue, 07 Jul 2026 00:00:00 -0000 It makes up less than 10% of the model — remove it and Claude can still talk, but its reasoning collapses to zero. Anthropic is already using it to catch fabricated data and spot when Claude senses it's being tested. It makes up less than 10% of the model — remove it and Claude can still talk, but its reasoning collapses to zero. Anthropic is already using it to catch fabricated data and spot when Claude senses it's being tested. 27:57 false https://best.xiaohu.ai/en/article/phosphor-dartmouth-quiz-study/ https://pic.xiaohu.ai/jiedu-podcasts/phosphor-dartmouth-quiz-study-en.mp3 Mon, 06 Jul 2026 00:00:00 -0000 In a 151-student trial, short-answer questions moved scores more than multiple choice, while almost no one touched the AI help sidebar In a 151-student trial, short-answer questions moved scores more than multiple choice, while almost no one touched the AI help sidebar 24:56 false https://best.xiaohu.ai/en/article/personalized-aac-app/ https://pic.xiaohu.ai/jiedu-podcasts/personalized-aac-app-en.mp3 Mon, 06 Jul 2026 00:00:00 -0000 No funding, no team — built purely to solve his own kid's problem. Clinics and schools started asking to use it anyway. No funding, no team — built purely to solve his own kid's problem. Clinics and schools started asking to use it anyway. 20:56 false https://best.xiaohu.ai/en/article/aiewf-loops-debate/ https://pic.xiaohu.ai/jiedu-podcasts/aiewf-loops-debate-en.mp3 Mon, 06 Jul 2026 00:00:00 -0000 A live audience vote couldn't be tallied because the venue lights were too bright to count hands, but a companion survey found 95% of teams already use agents while 59% worry about mounting technical debt. A live audience vote couldn't be tallied because the venue lights were too bright to count hands, but a companion survey found 95% of teams already use agents while 59% worry about mounting technical debt. 21:52 false https://best.xiaohu.ai/en/article/ai-schools-wealth-gap/ https://pic.xiaohu.ai/jiedu-podcasts/ai-schools-wealth-gap-en.mp3 Mon, 06 Jul 2026 00:00:00 -0000 While traditional schools are still figuring out AI, Silicon Valley and Wall Street families are already voting with their wallets. While traditional schools are still figuring out AI, Silicon Valley and Wall Street families are already voting with their wallets. 21:50 false https://best.xiaohu.ai/en/article/ai-junior-dev-collapse/ https://pic.xiaohu.ai/jiedu-podcasts/ai-junior-dev-collapse-en.mp3 Sun, 05 Jul 2026 00:00:00 -0000 US developer employment among 22-to-25-year-olds has fallen 19% in three years, even as new GitHub sign-ups hit their fastest growth ever. US developer employment among 22-to-25-year-olds has fallen 19% in three years, even as new GitHub sign-ups hit their fastest growth ever. 20:51 false https://best.xiaohu.ai/en/article/human-nature-persuasion/ https://pic.xiaohu.ai/jiedu-podcasts/human-nature-persuasion-en.mp3 Sun, 05 Jul 2026 00:00:00 -0000 Create psychological tension first, then offer a first step too small to refuse — it works for writing, selling, and job hunting alike. Create psychological tension first, then offer a first step too small to refuse — it works for writing, selling, and job hunting alike. 25:17 false https://best.xiaohu.ai/en/article/enterprise-ai-agent-guide/ https://pic.xiaohu.ai/jiedu-podcasts/enterprise-ai-agent-guide-en.mp3 Fri, 03 Jul 2026 00:00:00 -0000 Four implementation principles, backed by real-world data from L'Oréal, Lyft, and Rakuten Four implementation principles, backed by real-world data from L'Oréal, Lyft, and Rakuten 25:41 false https://best.xiaohu.ai/en/article/alibaba-page-agent/ https://pic.xiaohu.ai/jiedu-podcasts/alibaba-page-agent-en.mp3 Fri, 03 Jul 2026 00:00:00 -0000 MIT-licensed and model-agnostic — works with any OpenAI-compatible text model, though for now it only handles a single page view MIT-licensed and model-agnostic — works with any OpenAI-compatible text model, though for now it only handles a single page view 17:47 false https://best.xiaohu.ai/en/article/longcat-2-no-nvidia/ https://pic.xiaohu.ai/jiedu-podcasts/longcat-2-no-nvidia-en.mp3 Wed, 01 Jul 2026 00:00:00 -0000 Trained on 50,000+ domestic AI chips and 35 trillion tokens; most benchmarks come from Meituan's own evaluation framework, and the weights aren't truly open for download yet Trained on 50,000+ domestic AI chips and 35 trillion tokens; most benchmarks come from Meituan's own evaluation framework, and the weights aren't truly open for download yet 22:45 false https://best.xiaohu.ai/en/article/claude-science-workbench/ https://pic.xiaohu.ai/jiedu-podcasts/claude-science-workbench-en.mp3 Wed, 01 Jul 2026 00:00:00 -0000 Now in open beta. A coordinator agent marshals a team of expert agents to do the work, with a reviewer agent at the end dedicated to catching errors in citations and numbers — compute gets outsourced to AI, but raw data never leaves your local machine. Now in open beta. A coordinator agent marshals a team of expert agents to do the work, with a reviewer agent at the end dedicated to catching errors in citations and numbers — compute gets outsourced to AI, but raw data never leaves your local machine. 22:19 false