Good morning. It's Friday, February 6, and we're covering AI jobs debates, the stalled NVIDIA-OpenAI deal, Claude Opus 4.6 upgrades, and more.
New here? Keep up with the future of tech, sign up here. Have feedback? Send us a note: astro@forwardfuture.ai If you liked this email, share it with a friend.
✨ Claude Opus 4.6 Gets Smarter Upgrade
Anthropic released Claude Opus 4.6, upgrading its top model with stronger coding, reasoning, and long-running agent skills. It introduces a 1M-token context window in beta, enabling work across very large codebases and documents. The model leads key benchmarks for agentic coding and knowledge work. Anthropic says its safety performance matches or exceeds other models.
🚀 OpenAI Launches Frontier Platform
OpenAI unveiled Frontier, an enterprise platform for building and managing AI agents that do real work. It targets the gap between powerful models and messy deployment across organizations. Frontier adds shared context, permissions, feedback, and execution tools so agents can operate across systems. Early adopters include HP, Oracle, and State Farm.
👨💻 GPT-5.3-Codex Expands Agentic Work
OpenAI introduced GPT-5.3-Codex, its most capable agentic coding model, combining faster performance with stronger reasoning and professional skills. The model can handle long-running tasks involving research, tools, and full computer use, while remaining steerable during execution. It sets new highs on SWE-Bench Pro and Terminal-Bench while improving real-world productivity tasks.
🗣️ Voice Emerges As AI’s Next Interface
ElevenLabs CEO Mati Staniszewski says voice is becoming the primary way people will interact with AI as models grow more agentic. Speaking at Web Summit, he argued voice systems now combine emotional speech with reasoning, enabling more natural control of technology. The shift is driving investor interest, including ElevenLabs’ $500M raise at an $11B valuation.
🤔 FRIDAY FACTS
How Is a 47-Year-Old Spacecraft Still Talking to Earth?
Voyager 1 has been traveling through space for nearly 50 years, and it's still going strong. But what kind of tech is powering this interstellar spacecraft?
Stick around for the answer! 👇️
📽️ VIDEO
Use Clawdbot Like a PRO
Master Clawdbot: advanced setups, integrations, automation, model routing, and security tips to unlock a production-grade AI assistant.
📺️ FORWARD FUTURE LIVE
Today on Forward Future Live
👷♂️ LABOR
AI and Work: Job Loss Hasn’t Arrived, But Post-Labor Fears Are Growing
The New York Times published a discussion hosted by David Leonhardt on whether artificial intelligence is already reshaping employment—and whether it could eventually make human labor optional. Economists David Autor, Anton Korinek, and Natasha Sarin largely agree that current job-loss data is inconclusive, but sharply diverge on how seriously society should prepare for a future where AI substitutes for most work. The debate centers on whether history’s pattern of technological disruption creating new jobs will hold if artificial general intelligence (A.G.I.) removes labor as the economy’s central bottleneck.
This debate marks a shift in how mainstream economists talk about AI: not as a near-term job killer, but as a credible long-term challenge to the economic role of human labor itself. The disagreement isn’t about whether disruption is coming, but about timing—and whether governments should radically redesign income distribution before evidence forces their hand. → Read the full article here. (Paywall)
💰 FUNDING
$100Bn NVIDIA–OpenAI Deal Falters, Shaking Confidence in AI Funding Model
A widely reported $100bn investment and supply deal between NVIDIA and OpenAI, announced in September 2025, now appears unlikely to materialize. The arrangement was emblematic of “circular” AI funding, with NVIDIA money effectively returning via OpenAI’s purchases of its chips.
Both sides have since downplayed the figure, with NVIDIA CEO Jensen Huang stressing the talks were non-binding and OpenAI reportedly exploring alternative chip suppliers. The episode has unsettled investors, knocked roughly 10% off NVIDIA’s stock this week, and raised fresh doubts about who ultimately pays for AI’s capital-intensive expansion. → Read the full article here.
🛸 SPACE
Elon Musk Advances Plan for Orbital AI Data Centers
Elon Musk is moving from rhetoric to regulatory action on orbital AI data centers, following SpaceX’s FCC filing for a million-satellite computing network and its formal merger with xAI. The Federal Communications Commission has accepted the proposal for public comment, a key early step that FCC chair Brendan Carr publicly highlighted.
Musk argues that space-based solar power could make orbital data centers cheaper and easier to scale than terrestrial ones, predicting a tipping point as early as 2028. Skeptics note unresolved challenges, including hardware servicing and the fact that energy is only one of many data-center costs. → Read the full article here.
🛰️ NEWS
What Else is Happening
🔮 Luckey Sees $1,000 Trucks: Anduril cofounder Palmer Luckey says AI could drive manufacturing costs near zero, making vehicles as cheap as a Ford F-150 for $1,000.
🥑 Meta Touts ‘Avocado’ AI Model: An internal memo says Meta’s new model, code-named Avocado, is its most capable yet, signaling an aggressive push to close the AI performance gap.
🎮 Take-Two Embraces AI, Not GTA 6: CEO Strauss Zelnick says the publisher is testing generative AI companywide, but insists it plays zero role in Rockstar’s GTA 6.
📲 Meta Spins Out Vibes App: Meta is testing a standalone app for its AI-generated Vibes videos, aiming to rival OpenAI’s Sora and deepen engagement around AI-only short-form video.
📈 Gemini Powers Google’s AI Rebound: Alphabet posted a breakout quarter as Gemini drove growth across search, cloud, and YouTube, easing fears over Google’s soaring AI spending.
🗒️ FEEDBACK
Help Us Get Better
What did you think of today's newsletter? |
🤔 FRIDAY FACTS
Voyager 1 Is Running on 68 Kilobytes of Memory
That’s not a typo. Voyager 1, now over 15 billion miles from Earth, operates on just 68 KB of memory—less than a single low-res photo.
So how is it still working?
Tightly written code that wastes nothing
Radiation-hardened hardware that laughs at Jupiter’s fury
Flawless remote updates, despite a 22-hour signal delay
And the signal we get? It’s about 0.1 billion-billionth of a watt. Basically: a whisper from the edge of the solar system. Sometimes, old tech flies circles around the new.
