What I do as an Operations Research: professional:
I figure out the best thing to do when you don't know what the hell's going on, and when you think you do but you really don't.
Here is ChatGPT 5.5 Pro Extended thinking mode being blunt ... in its disappointment with Codex 5.5 xhigh.
"My assessment is blunt: V4.33B made a useful audit pass and confirmed the near-miss state, but it did not make enough new engineering progress."
Does @sama realize that ChatGPT 5.5 Pro extended thinking mode is much smarter and more capable in math and algorithms than Codex 5.5 xhigh? Why is there no "Pro" version of Codex? The multiple referees I have created to review Codex's work before it can complete are a dismal failure - the referees are passing work which should flunk per their mandates. Did whatever post-training was done to make Codex good at coding dumb down its math and algorithm abilities relative to ChatGPT 5.5 Pro? Perhaps that will get resolved when ChatGPT and Codex are merged into one, hopefully without dumbing down the merged entity's math and algorithm abilities?
@RicursiveAI@eetimes@sallywf Amdahl's law will soon make tapeout the pain point in chip design, verification, and manufacture..
Will there be a phase 4, greatly automating tapeout, reducing time and cost to manufacture?
Bottom line, TSLA is overvalued like crazy unless it comes through with huge revenues and profits from new businesses, such as robots and AI, which are not reflected in its current sales, and therefore in its Price to Sales ratio. TSLA stock is as high as it is because there are a lot of Elon believers. The market doesn’t always get it right.
A couple f decades ago, Boeing financed a lot of the planes it was “selling”. When some of the buyers went bankrupt, Boeing took a bath. A lot of its revenue had been illusory.
Anyhow,If the market believed Nvidia is worthless, its market cap would be considerably below $5.2 trillion, the highest in the world. And not everyone is sure those other companies you mentioned are a sure thing. If you believe strongly enough, you should leverage to the max, buying NVDA, and shorting TSLA, and by next month, perhaps also shorting SPCX, because it’s not easy to lose as much as money as SpaceX. So SPCX must have a very special kind of hype, and you can prepare for a possible TSLA SPCX merger.
If you own stock on margin, and it goes down, and then back up 66 2/3% (so back to original stock price), then not even counting borrowing cost, you might be down due to margin calls forcing selling at or near the low, and then not having full participation on the way back up. Path dependent returns can be a bitch when you buy (own) on margin. Up then down is better than down then up when both paths arrive at the same end point.
@JFPuget @slopwareindy @notdrvx xhigh all the way, baby. And even then, Codex might look for any excuse or technicality to minimize its actual work performed.
@deedydas Ricursive Intelligence raised at $750 million valuation in seed round in Dec 2025. Series A in Jan 2026 at $4 billion valuation. So the seed round was less than $1 billion.
@mmaaz_98 I'll counterpropose with Von Neumann, because he presented the complete duality theory of Linear Programming to Dantzig when Dantzig told him what he had done on Linear Programming. https://t.co/jeKx7dfx9j
@TheSIAMNews Congratulations to all new Fellows.
But as usual, almost all the Fellows are academics.
Either remove Industrial from SIAM's name, or make SIAM more industrial.