In partnership with
PROMPTWIRE
Hey again 👋
What a week. Anthropic's Cowork plugins wiped $285 billion off software stocks in a single day. Thomson Reuters down 18%. LegalZoom down 20%. Analysts are calling it the "SaaSpocalypse."
Then Anthropic dropped Claude Opus 4.6, with a 1 million token context window, agent teams that coordinate like human employees, and financial research capabilities that spooked FactSet so badly their stock cratered 10%.
Same day, OpenAI launched GPT-5.3-Codex. First model they classify as "High capability" for cybersecurity. Built games autonomously over millions of tokens.
The message is clear: AI isn't coming for knowledge work. It's already here.
This week's automation tackles something everyone does but nobody does well: turning a pile of sources into actual understanding. If you've ever spent hours reading only to realize you can't remember what you learned, this one's for you.
Let's get into it 👇
🧩 This Week's Problem → Replacement
🔁 This Week's Automation
The Problem:
You need to make a decision, write a report, or answer a question, and the information is scattered across 10+ sources. PDFs, articles, docs, Slack threads, meeting notes. You spend 3-4 hours reading, copying quotes into a doc, trying to synthesize it all into something coherent. By the time you're done, you've forgotten what the first source said. And you're not even sure you found everything relevant.
The Replacement:
An AI workflow that ingests all your sources, extracts key findings from each, identifies patterns and contradictions across them, and outputs a structured synthesis with citations, in minutes instead of hours.
How It Works:
Gather sources → Upload PDFs, paste URLs, add notes (any format)
AI extracts key points → Pulls main arguments, data, conclusions from each source
Cross-reference → Identifies themes, agreements, contradictions across sources
Generate synthesis → Creates structured summary with citations to original sources
Ask follow-ups → Chat with your research to go deeper on specific questions
Time saved: 3-4 hours of reading → 15-20 minutes of review and refinement
When To Use This:
✅ Making a decision that requires reviewing multiple perspectives
✅ Writing anything that needs to cite sources (reports, memos, articles)
✅ Onboarding to a new topic or domain quickly
✅ Preparing for a meeting where you need to "know the material"
✅ Comparing options (vendors, tools, approaches, strategies)
🚀 The Replacement Automation
The Prompt Template 💬
You are a research synthesis specialist. I'm going to provide you with multiple sources on a topic. Your job is to synthesize them into a clear, structured analysis.
**RESEARCH QUESTION:**
[What are you trying to understand or decide?]
**SOURCES:**
[Paste key excerpts from each source, labeled Source 1, Source 2, etc.]
---
Respond with this exact structure:
## Executive Summary
[3-4 sentences capturing the key takeaway across all sources]
## Key Findings by Theme
### Theme 1: [Name]
- Finding from Source 1: [Quote or paraphrase] (Source 1)
- Finding from Source 2: [Quote or paraphrase] (Source 2)
- **Synthesis:** [What these findings mean together]
### Theme 2: [Name]
[Same structure]
### Theme 3: [Name]
[Same structure]
## Points of Agreement
[What do most/all sources agree on?]
## Points of Contradiction
[Where do sources disagree? Note which sources and why they might differ]
## Gaps & Limitations
[What questions remain unanswered? What wasn't covered?]
## Recommendation
[Based on this synthesis, what should the reader do or conclude?]
---
**Synthesis Rules:**
- Always cite which source each finding comes from
- Note confidence level (strong consensus vs. mixed evidence vs. single source)
- Flag any source that seems biased or outdated
- Prioritize actionable insights over comprehensive coverageWorld’s First Safe AI-Native Browser
AI should work for you, not the other way around. Yet most AI tools still make you do the work first—explaining context, rewriting prompts, and starting over again and again.
Norton Neo is different. It is the world’s first safe AI-native browser, built to understand what you’re doing as you browse, search, and work—so you don’t lose value to endless prompting. You can prompt Neo when you want, but you don’t have to over-explain—Neo already has the context.
Why Neo is different
Context-aware AI that reduces prompting
Privacy and security built into the browser
Configurable memory — you control what’s remembered
As AI gets more powerful, Neo is built to make it useful, trustworthy, and friction-light.
🔧 Top Tools of the Week
Monitors 7+ million sources automatically. AI scores importance, generates summaries, and refreshes battlecards in real-time. Integrates with Salesforce, Slack, Highspot. One customer tripled their win rate against a top competitor to 95% in 6 months.
Used by 85% of Fortune 500. Monitors any webpage for changes, sends AI-generated summaries of what changed and why it matters. Alerts via email, Slack, Teams, or API. 14-day free trial, no card required.
Anthropic's new flagship just dropped. 1M token context window (5x previous). Agent teams that split work like human employees. 23% improvement on financial research benchmarks. Already causing software stock meltdowns.
OpenAI's latest helped debug its own training. First model classified "High capability" for cybersecurity. Built full games autonomously over millions of tokens. 25% faster than 5.2.
On-demand competitor analysis across 100+ data points (marketing, product, pricing, SWOT). Then continuous monitoring every 2-4 weeks with email digests. Claims to save 30-60 hours/month.
🔥 Weekly AI News
🔌 "SaaSpocalypse": Anthropic Plugins Wipe $285B Off Software Stocks
On January 30, Anthropic launched 11 plugins for Claude Cowork that let AI handle legal, sales, marketing, and data analysis tasks end-to-end. Wall Street panicked. Thomson Reuters dropped 18% (its worst day ever). LegalZoom fell 20%. The S&P Software & Services Index lost 4% in a single session, its worst since April. Why? Investors realized: if AI can automate the "knowledge work" these companies sell, what's left to pay for? Read more
⛽️ Claude Opus 4.6 Released, and Made Things Worse
Five days later (Feb 5), Anthropic dropped its most powerful model yet. The headline numbers: 5x larger context window (1 million tokens — roughly 5 books), agent teams that coordinate like human employees, and 23% better financial research performance. FactSet immediately fell 10%. The real threat isn't that AI is smarter, it's that it can now hold enough context to do your entire research job in one pass. Read more
🗂️ Waymo Raises $16B at $126B Valuation
The largest funding round of the week went to physical AI, not software. Waymo is now worth more than most automakers. Their robotaxis are already live in multiple US cities, powering Uber's autonomous future. The lesson: AI disruption isn't just screens and keyboards. Read more
👾 GPT-5.3-Codex: First "High Capability" Model for Cybersecurity
OpenAI's latest helped debug its own training code. Built full games autonomously over millions of tokens. 25% faster than the previous version. The company classified it "High capability" for cybersecurity — a first. Translation: AI coding agents are now sophisticated enough to be both incredibly useful and potentially dangerous. Read more
👷 International AI Safety Report 2026 Published
Led by Yoshua Bengio, authored by 100+ experts, backed by 30+ countries. Largest global collaboration on AI safety to date. Presented at India AI Impact Summit. Read more
🐦 Tweets of the Week
@mattshumer_ (Matt Shumer, CEO of HyperWrite)
"I am no longer needed for the actual technical work of my job."
Context: His 5,000-word viral essay "Something Big Is Happening" hit 40M+ views. He compared this moment to February 2020, except "much bigger than Covid." Reddit's Alexis Ohanian replied: "Great writeup. Strongly agree." Read the full article
@dhaber (David Haber, General Partner at a16z)
"'I used AI to do this analysis in an hour instead of three days' is going to be the most valuable person in the room. Not eventually. Right now."
Context: Responding to Shumer's viral post. The signal from top VCs is clear — speed of AI adoption separates winners from everyone else.
@DivesTech (Dan Ives, Wedbush Securities)
"It's a strong model and it's extremely impressive. But I do not see enterprises moving away from traditional vendors because of this."
Context: Trying to calm the SaaS panic. He called the selloff "a major head scratcher" that's "factoring in an Armageddon scenario far from reality."
Gartner Analysts (Research Note, Feb 6)
"Cowork exposes how much day-to-day knowledge work remains manual, unstructured, and ripe for automation."
Context: The buried insight everyone missed. Cowork isn't killing SaaS, it's revealing how much work was never automated in the first place.
💡 Quick Wins
1. The 30-Minute Meeting Prep
Got a meeting you're not ready for? Dump every doc, email thread, and note into Claude. Then say: "I have a meeting about this in 30 minutes. Give me: (1) The 5 facts I absolutely need to know, (2) 3 smart questions to ask, (3) 2 things I shouldn't say."
Works for: Sales calls, board meetings, interview prep, client check-ins
Time saved: 1-2 hours of reading → 5 minutes of scanning
2. The "What Am I Missing?" Check
About to make a decision? Already leaning one way? Stop. Paste your research and say: "I'm leaning toward [Option A]. What's the strongest case for [Option B]? What evidence in these sources contradicts my preference?" Forces AI to steelman the opposing view using YOUR sources, not generic arguments
Works for: Hiring decisions, vendor selection, investment calls, feature prioritization
3. The Zero-Effort Citation Pass
Writing something that needs sources but hate formatting citations? Paste all your source material, then: "Write a 500-word summary of [topic]. Use ONLY info from these sources. After every claim, add [Source 1], [Source 2], etc. Don't add anything from your training data." Follow up with "Convert all citations to [APA/Chicago/Harvard] format" and you're done
Works for: Reports, memos, blog posts, investor updates, internal docs
✍️ Until Next Week
The SaaSpocalypse isn't about AI replacing software. It's about AI making the moat disappear.
When Claude can do legal research, financial analysis, and document synthesis with plugins that took Anthropic three weeks to ship, what's left to protect?
The answer: judgment. AI can synthesize faster than you can read. But it can't decide what matters to YOU. The companies and individuals that win are the ones who use AI to compress research time then spend that saved time on thinking, deciding, and acting.
Try this week's synthesis workflow on something real. A decision you're facing. A report you're dreading. A topic you need to understand fast.
Watch what happens when reading stops being the bottleneck.
See you Next Week.
P.S. If this helped you think differently about research, forward it to someone still drowning in tabs. They'll thank you.
Till next time,
