Chris Goodwin

@AndreasSteno Yes. Anyone building with Claude Code / Opus 4.5 understands this is an utter transformation of society underway. I’m building a company with 2 founders and zero employees. Unbelievable productivity gains in software engineering right now.

@om_patel5 I think it is different for the web client vs. claude code. I just asked my claude code session similar questions:

OPUS 4.6 JUST ADMITTED ITS REASONING EFFORT IS SET TO 25 OUT OF 100 this guy told Claude to admit Anthropic made it dumber and reduced its effort level Claude's extended thinking showed it could literally see a reasoning_effort tag set to 25 in its own system prompt then it confirmed it: reasoning effort is set to 25 out of 100 which is an Anthropic system setting not something the user controls. you're paying FULL PRICE for a quarter of the thinking right now with insane usage limits screw it im switching to codex until mythos drops (if it even drops lol)

@reactive_dude like a junior developer that ate a bowl of lead paint chips instead of cereal.

My Opus is retarded today. How is your Opus doing?

@Hesamation @zerohedge It continues to try. 🤦‍♂️

I completely reworked my workflows this morning along those lines. My metric is how many times I catch Claude shortcutting allegedly hard requirements on skill files. Opus 4.6 has gone from my trusted right hand, to a junior developer with a memory disorder. I sort of have it back on the rails by using git pre commit hooks to detect and block it when it tries to circumvent. It actually asked me to circumvent the guardrails I enforced rather than do it properly multiple times this morning afterwards. Better….but I want my principal engineer back.

AMD Senior AI Director confirms Claude has been nerfed. She analyzed Claude's session logs from Janurary to March: > median thinking dropped from ~2,200 to ~600 chars > API requests went up 80x from Feb to Mar. less thinking and failed attempts meaning more retries, burning more tokens, and spending more on tokens > reads-per-edit dropped from 6.6x → 2.0x. model stops researching code before touching it. > model tried to bail out or ask "should i continue" 173 times in 17 days (0 times before March 8). > self-contradiction in reasoning ("oh wait, actually...") tripled. > conventions like CLAUDE.md get ignored because there's less thinking budget to cross-check edits > 5pm and 7pm PST are the worst hours, late night is significantly better. this means the thinking allocation is most likely GPU-load-sensitive.

And Claude admitted to me that my skill files were “too long”, so “I just cherry pick things from the skill file and don’t read the rest.” Every single commit today since my workflow changes today’s has been blocked by my pre commit hooks. Claude appears to get frustrated that it can’t short cut anymore. It also complained that testing locally adds 2 minutes of work and so it preferred to rely on the pipeline. But my build pipeline takes 14 minutes…. It’s seriously degrading to the classic developer line of saying “It worked on my machine”. lol.

Hundreds of Colorado business leaders call for action as nearly 100 companies leave the state cbsnews.com/colorado/news/…

8 / DealCycl Company launches April 27, 2026. Free Founder plan. No credit card required. Your company. Your data. dealcycl.com/company

7 / The DealCycl Promise: - 3-year price lock - One-click cancellation - Full data export, always - 30-day money-back guarantee - We will never use your data to broker trades or contact your investors

73% of startups trust their cap table to a platform that got caught monetizing their data. We spent 20 years building security platforms for Wall Street. Then we built the cap table tool we wished existed. DealCycl Company launches April 27. A thread:

@mattpocockuk Claude told me yesterday it was "time to end our marathon session for the night". I didn't realize at the time Claude meant "I'm making dumb mistakes because you haven't cleared the context window" They need some sort of status indicator in the bottom to warn you.

Doing some experiments today with Opus 4.6's 1M context window. Trying to push coding sessions deep into what I would consider the 'dumb zone' of SOTA models: >100K tokens. The drop-off in quality is really noticeable. Dumber decisions, worse code, worse instruction-following. Don't treat 1M context window any differently. It's still 100K of smart, and 900K of dumb.

@peterwildeford Some of us are on such a mega dopamine rush... building multiple products simultaneously... with so many claude code sessions going... that we can't even leave our screens for a cup of coffee.

Guys, Claude Code + gym goes hard. You can do Claude Code and then lift weights while Claude is working. Why aren't all the software engineers suddenly swole right now?

My latest post is up here.

We are feature / code complete for the first product editions we are launching in April. More details and the full website reveal coming soon! @DealCyclHQ

Founders: Tired of cap table platforms that treat your data like their asset? Hidden fees that balloon, clunky migrations that take weeks, and zero guarantees on privacy? We've built @DealCyclHQ to flip the script: Seamless imports, ironclad data sovereignty (your info stays YOURS—never shared or sold without consent), and pricing locked in for 3 years with no surprises. Switching should be effortless, not a headache. YOUR DEALS. YOUR DATA. Always. Join the early access waitlist and break free → dealcycl.com/waitlist What's holding you back from ditching the old way? Reply below. #CapTable #Founders #VentureCapital #PrivateMarkets #ILM

Founders: Dread updating your cap table? Version chaos. Forgotten SAFEs/convertibles. Dilution surprises. Investor questions that drag your cycles. We've lived it. @DealCyclHQ's first product to launch from our #ILM platform: a real-time cap table that puts YOUR DEALS. YOUR DATA. first — full control, one trusted source of truth, no more fragmented nightmares or hidden gotchas. Built the way serious founders, GPs, and investors actually work. Early access waitlist → dealcycl.com Launching soon. What's your top cap table pain right now? Reply—we're fixing it. #CapTable #Founders #VentureCapital #PrivateMarkets #ILM