(FORTUNE Magazine) – There is little about Wall Street that still feels larger than life. The scandals of the last bull market have smudged much of the luster, and consolidation has dulled some of the history, making the mythic Street seem sordidly human. But if there's one place the old mystique lingers, it's Goldman Sachs. Its storied past is one reason. For years, to be a partner at Goldman was to be a prince on Wall Street; the firm's IPO five years ago was wealth creation on a magnificent scale. Goldman Sachs is the only one of the Street's giants to have survived for 135 years, untouched by megamerger or financial catastrophe. And today it is arguably more powerful than ever.

Just listen to how top businesspeople describe the firm: Jim Coulter, a founding partner of the Texas Pacific Group, calls it "one of the great talent networks, not just on Wall Street but in industry." Jack Levy, who joined Goldman in 2000 from Merrill Lynch, where he ran its merger department, has long marveled at its "mysterious force." Bill George, the former CEO of Medtronic who joined Goldman's board in 2002, says a friend once said that he never, ever thought George would become a director at an investment bank. George's response: "I didn't. I joined the board of Goldman Sachs."

Goldman's financial performance speaks for itself. Despite the bad economy, despite the fact that in the business most people associate with Goldman--investment banking--profits have plummeted in three short years from $1.7 billion to just $207 million, despite being in an industry that lacks "recurring revenues," Goldman has continued to make money with money. Since 1996 the firm's equity base has grown at an average annual rate of 22%, while revenues have grown at almost 15%, vs. 8% for peers, according to Bank of America. Goldman's return on equity is well above 20%, and in the first quarter of 2004 it had its most profitable quarter ever, earning almost $2 billion before taxes.

And yet Goldman's stature far outstrips its actual size. The firm is dwarfed by competitors. Its market capitalization of $42 billion is roughly one-sixth that of Citigroup's $238 billion. Does its stature outstrip reality too? Goldman, as more than one person will tell you, is the firm that others on the Street "love to hate." You hear constant sniping that the glory days are gone, that Goldman doesn't deserve its aura, and even that CEO Henry "Hank" Paulson will be the last CEO of Goldman Sachs--because the firm will be sold on his watch.

There are, in fact, legitimate questions about its future. Goldman was a private partnership for 130 years. It has been a public company for only five years. In a world where investors are crying for increased transparency, Goldman is opaque. There's no way for any outsider to know exactly how the firm makes its money. And that's scary, because there's huge risk associated with those huge profits. What's the guarantee that Goldman will continue to draw the winning hand? The questions are reflected in the stock, which now sells for about $85, less than two times book value and at the bottom of its historical trading range.

Mystique also implies mystery, so it's fitting that one of the words often used to describe this Wall Street brokerage is "secretive." Goldman isn't just mum about how it makes money. The firm would prefer to keep all its secrets. But it has faced increasing pressure to open up. And when it finally made the decision to grant access to FORTUNE, in what is typical Goldman behavior, the whole firm got behind it. (It was almost difficult to make them go away.) Goldman, though, turns out to be every bit as fascinating as its reputation suggests. For all its monolithic image, this is a place that's full of surprises.

Start with the physical surroundings of the firm. There's nothing that shouts either "Goldman Sachs" or "money." The drab brown headquarters at 85 Broad Street famously bears no sign. Inside, you're struck not by elegance, and certainly not by opulence, but rather by the shabby state of affairs--the faded, torn carpets, the grungy cubicles, the utterly ordinary offices. Treasurer Elizabeth Beshel, who was a junior analyst back in 1990, points out that a giant ink stain she left on her old cubicle is still there. (The firm is building a new headquarters in downtown Manhattan.) Nor will you spot a lot of flashy investment bankers. Jon Winkelried, who is the co-head of Goldman's massive fixed-income, currency, and commodities division, sports a shirt with a torn sleeve and missing buttons.

Then, under the immutable surface of the Goldman name, there's the massive--if that's a big enough word--change at the firm since its May 1999 IPO. Head count almost doubled, from 13,000 to more than 25,000 at the peak; since then, Goldman has undergone wrenching layoffs, second only to Merrill Lynch in percentage terms. Departments like investment banking were decimated, with roughly 40% of the staff gone at the bottom. Here's the shocking reason: Banking now represents a mere 5% of Goldman's profits, down from 34% in 2000. Although Goldman is still the worldwide market share leader in mergers and acquisitions, that department, once an elite unit within an already elite firm, has been disbanded due to lack of business; M&A bankers now work in industry groups. Even so, Goldman's size seems almost unwieldy. Today it employs 20,000 people--roughly half of whom have been with the firm for less than four years.

And then, there's the surprise of Goldman's top people--they have power, money, and prestige. You'd think they'd feel untouchable. But this is part of the contradiction that is Goldman: While there is plenty of arrogance at this firm, few of its executives seem to take anything for granted. Gary Cohn, who co-heads Goldman's equity business, says, "We don't create something people have to buy to survive or something uniquely different." Indeed, Goldman has no factories, no monopoly. Quite the opposite: The firm faces ferocious competition in every single one of its businesses, with no guaranteed edge. Cohn recounts a conversation he had with legendary Goldman trader Mark Winkelman back in the mid-1990s: "There's no difference between us and every other investment bank out there --except your hard work, your management, and the people you hire," Winkelman told him. "Think about it. We work in the same buildings, we have the same computers, we fly on the same planes, we sleep in the same hotels. We even have the same clients." Says Cohn: "I think about this ten times a day."

The culture in which people like Cohn grow up is not a warm and nurturing one. The men at the top of Goldman Sachs (and they are almost all men) seem relatively normal. You might even want to have a drink with some of them. But just under the surface--and in some cases not even under the surface--is a bulldog intensity. It's an intensity that is born of ambition, greed, paranoia--and believe it or not, a sense that the rest of the world has the advantage. You could even say that Goldman, as a firm, has a collective chip on its shoulder.

One explanation is that it is not the "white shoe" firm of lore. The legendary Sidney Weinberg, who was Goldman's chairman from 1930 to 1969, wasn't a guy who was born rich but rather one who began by shining the shoes of partners. And the firm's current partners are still the guys who claw their way up. Paulson, who grew up on an Illinois farm, says that he has a hard time thinking of anyone today who has done well at Goldman who came from a wealthy background. There are a few, of course, whom Paulson is missing, but the point is well taken. Cohn, who joined Goldman in 1983 from the floor of the New York Commodities Exchange, where he worked for himself as a silver trader, says he's so dyslexic that he struggled to read until the eighth grade. "We were all on the wrong side of everything," he says. "Years of being told you were going to fail makes you say, 'I'll show you.'"

Then there's the fear that everyone wants what Goldman has. "We live day in and day out with the understanding that our business is constantly under attack," says Mike Evans, who co-heads the equities business with Cohn (and who, as a member of the Canadian rowing team, won a gold medal in the 1984 Olympics). "If we don't constantly adapt and look for new ways to make money, we'll be No. 2, then No. 3, then No. 7 ..."

Then there's the internal competition. For a Goldman executive, it's unclear whether the worst thing about losing a piece of business is losing it--or confessing the loss to others at the firm. At other banks it's enough to be a good manager. Not at Goldman: You'd better also make money for the firm. Each year Goldman ranks its employees into four quartiles, and you will hear people dismissed as "fourth-quartilers." "This is an intensely competitive place, even more so now that the stakes have gotten bigger," says Kevin Kennedy, who has spent the past three decades at Goldman and is now head of what the firm calls, without irony, Human Capital Management. "There are plenty of external factors, but a lot of it is self-induced."

Goldman is, after all, a ruthless place, one that sucks very smart people in and spits them out with surprising rapidity. The firm's executives all say they worry about losing people, these days particularly to hedge funds. But for now there are always more smart youngsters desperate for their chance. "It's like the Zulus," says Paul Deighton, chief operating officer of the firm's European business, referring to fierce, skilled warriors. "There's always another one coming over the hillside to fight."

You see that ruthlessness in the short tenures of most Goldman partners. (Despite the IPO, partners are still elected at Goldman every two years. There are 253 people today who get a salary of $600,000 per year and share in the firm's profits.) It's up or out, and if you're not getting ahead, you're falling behind. The average tenure of a partner was just ten years back in the 1980s; today it's eight years. More than half of the partners at the time of the IPO have left. They leave because they're rich enough--the pre-IPO partners, as they're known at Goldman, each got stock worth an average of over $100 million (at today's prices)--they're tired enough, or because they were asked to leave. "You don't sign on to Goldman for the retirement program," says one former partner. "You have a phase of the moon to make your money, and then you have to make room."

All this seems to make Goldman people work just a little bit harder. Clients get called before anyone else on the Street has thought to make a call. Jeff Immelt, GE's CEO, says that John Weinberg, Sidney's nephew and one of three heads of investment banking at Goldman, was one of the few bankers on Wall Street who nurtured a relationship with Immelt when Jack Welch was running GE. "It has served him well," Immelt says. (Goldman, along with Morgan Stanley, led the almost $3 billion initial public offering of GE's Genworth insurance business earlier this year.)

But what really makes Goldman Goldman is that the firm, over its century-plus history, has managed to harness all the individual greed, ambition, and paranoia into genuine teamwork. It's "we" rather than "I." There are few "heads" of anything at Goldman. They're all co-heads. So there is little of the usual corporate sanctimony in Goldman's use of the phrase "culture carriers" to describe Cohn and other top executives, many of whom have been at the company for at least two decades, and who live and breathe Goldman. (The phrase sounds as if it refers to bearers of infectious disease, but it means those who transmit the Goldman soul to future leaders.) It's real because it reflects pure survival instinct.

And while there's no guarantee of permanence, there is something self-perpetuating about all this. "Goldman has done well because it attracts good people, and good people attract other good people, and in that business, that's the stock in trade," says Immelt, who does not play favorites with Wall Street firms. That special edge means that Goldman doesn't need to indulge its high-priced talent as much as rival firms do theirs. Remarkably, Goldman does not give bankers multiyear salary guarantees and does not pay traders a percentage of their earnings.

You would think that such a culture requires an inspirational leader to keep it humming. But Goldman doesn't have one. CEO Paulson once told his alumni magazine, "I'm not an inspirational leader," and he wasn't lying. The chief executive of Wall Street's most powerful firm is awkward rather than glib--and a study in surprises, if not contradictions. Although he owns over $350 million of Goldman stock, he is a Christian Scientist who has inveighed against conspicuous consumption. He has never sold a Goldman share except to donate money to charity, and his worst vice seems to be too many Diet Cokes. "Hank doesn't drink, doesn't smoke, is still on his first wife, and you'd never catch him dead in a Porsche," says a former Goldman partner. Inside the firm, Paulson is seen as smart but not cerebral, prone to action rather than complicated thoughts. And he's incredibly direct. Paulson himself says that one of his big challenges as CEO has been not going "brain to mouth."

On a recent Saturday in July, Paulson was not out in the Hamptons, but rather about a half-hour from Chicago's O'Hare airport, in the outskirts of Barrington, Ill. Off a back road, there's a small gray sign that says MERIMAR FARM and, under that, H.M. PAULSON. The first house along the winding driveway belongs to Paulson's 81-year-old mother, Marianna. A bit farther, on five acres that Paulson's dad gave him, is the modest glass-and-wood house that Paulson and his wife Wendy began building in 1974, when, after two years of staff jobs in the Nixon administration, he joined Goldman as an investment banker in the Chicago office earning $30,000 a year. The Paulsons still spend many weekends here. Lunch is blueberries, peaches, and fresh vegetables from the garden that Marianna tends herself, and the conversation is mainly about the 58-year-old Paulson's twin passions of nature conservancy and fly-fishing. After lunch, Paulson, a 6-foot-1 former Dartmouth football star who is wearing shorts and a polo shirt, kicks back on the porch, his feet up on a bench.

But beware the impression of the CEO at rest. Paulson has hard, carved features without a hint of softness. When we call from the airport to say we're running a bit late, his response, in his gravelly voice, is, "Drive fast." He doesn't like to waste time, and he doesn't like to lose, especially when he's fighting for a piece of business. He's a "really, really ruthless competitor," says one client. "And he knows how to wield power." Paulson, after all, was the leader of the coup that ousted his co-CEO, Jon Corzine (now a U.S. Senator from New Jersey), back in 1999, and he's been schooling himself in how to be a good CEO since his early days as a banker, when he chose people like John Bryan, the former CEO of Sara Lee, and James Johnson, the former CEO of Fannie Mae, as mentors. (Both now sit on Goldman's board.) In early 2002 there was a well-publicized incident at a New York investment conference where Paulson said, "I don't want to sound heartless, but I'm going to tell you the facts. In almost every one of our businesses there are 15% to 20% of the people who add 80% of the value ... you can cut a fair amount and not get into muscle." (He later apologized for the remark.)

Paulson's schedule is frenetic--dinner with China's Premier one day and lunch with Germany's Chancellor the next--and he spends roughly one-third of his time on the road, visiting clients. They notice. "He is extremely aggressive from the standpoint of acting out of sequence," says Immelt, meaning that Paulson, unlike others, visits not just when deals are happening but also when they aren't. Texas Pacific's Coulter dates the improvement in his firm's relationship with Goldman to a visit Paulson made five years ago. "They show up," Coulter says of Paulson's visits. "I've never asked, and they show up. You do not often see that in other CEOs of his stature." Says Frank Dangeard, a former senior executive of France Telecom: "He has time to give his clients calls, and not many chairmen of Wall Street firms do." Paulson has been to China 68 times since 1992 and is closer to China's leaders than many in the U.S. government.

It's quite clear that Paulson has no plans to relinquish his power. Paulson says that he "wouldn't put a time limit" on his tenure at Goldman, and that he has no intention of following in the Washington-bound footsteps of former Goldman CEOs like Bob Rubin, Steve Friedman, and Corzine.

His tight grip on power has been a cause of no small degree of upheaval. The two heirs apparent to Paulson were the fiercely ambitious John Thornton, a veteran M&A banker who spearheaded Goldman's push into Europe and Asia, and chief operating officer John Thain. Each had spent more than two decades at Goldman and shocked outsiders with their abrupt departures last year. First came Thornton, who now teaches at Beijing's Tsinghua University. And then, in late 2003, Thain left to take over the New York Stock Exchange.

Inside Goldman, however, it was clear to some that Thain and Thornton had already been supplanted. In 2002 an executive named Lloyd Blankfein, who had made his name running the firm's powerhouse fixed-income, commodities, and currency trading business, or FICC, made more than Paulson--$12.7 million vs. $9.6 million. In early 2003, Blankfein became the first and only employee since the IPO to be named to Goldman's board. By last fall, people who were seen as "Lloyd people," such as Gary Cohn, were running many businesses within Goldman. And in December 2003, Blankfein became president and COO.

It's tempting to see something Machiavellian in this. But the real story is beautifully, if brutally, simple: In the past few years FICC has been a major source of the firm's profits.

Blankfein's story is the extreme version of the Goldman Sachs narrative that started with Sidney Weinberg. He grew up in a tough neighborhood in Brooklyn, was the first in his family to go to college, and put himself through Harvard and Harvard Law. He joined not Goldman Sachs but a rough-and-tumble commodities trading business called J. Aron, which Goldman acquired at the peak of the commodities boom in 1982, right before Blankfein's arrival. The business promptly fell off a cliff, and Blankfein, who started as a gold salesman, played a key role in rebuilding it, taking it from its roots in coffee and metals trading into oil and foreign exchange. J. Aron was a stepchild at Goldman Sachs--J. Aron partners were amazed when Goldman executives even knew their names--but by the early 1990s it accounted for one-third or more of the firm's profits. And that made it a powerful perch. In 1997 the firm merged J. Aron with Goldman's fixed-income business to create FICC, which was run by Blankfein. Today FICC makes money buying and selling huge quantities of everything from mortgages to oil to euros and much, much more.

The first thing everyone says about Blankfein is that he's smart, or even "wickedly smart." He has a temper, and he's got a kind of intellectual charisma that Paulson lacks. One partner, who says there's a "sun god" phenomenon about Blankfein, was amused to see FICC traders, testosterone up to their eyeballs, sitting around in a conference room obediently awaiting Blankfein's arrival before starting their meeting. He is not physically prepossessing--in the past few years he's lost 50 pounds, shaved his beard, and quit smoking. (Like Paulson and many of Goldman's top executives, Blankfein is bald.) He puts his feet up on the table in the same pose of pseudo-relaxation as Paulson. But in Blankfein's case, you might just be lulled into relaxing. He's gregarious, quick with a punch line, and charmingly self-deprecating. "Who knows?" he shrugs when you ask about his career. "I haven't ever been in the position where I can say, 'Gee, I won't come in tomorrow.'" Except you know that there's another train of thought--or two, or three--taking place in his brain.

One way to look at the shifts in the 30th-floor executive suite, where Goldman's head honchos sit in glass-walled offices with views of the East River, is as evidence of a culture in flux, of a power shift from Goldman's bankers to its traders. But you can also view Blankfein's rise as proof that Goldman is, at least for now, still Goldman. Throughout the firm's history, he who is successful, be he a banker or a trader or both at the same time, has risen. (Bob Rubin made his name running the arbitrage business, in which the firm bet its own capital on takeovers; his co-CEO, Steve Friedman, helped build the M&A business.) Ultimately, Goldman is agnostic: It doesn't have to be a banking firm or a trading firm. Instead, Goldman seems to have a gut ability to go where the money is--driven, perhaps, by the horror of being where it's not. "There is nothing worse than being at Goldman Sachs when we're not making money," says Kennedy. "Nothing."

Today Goldman is clearly making money. "How?" and "Will it continue?" are the questions many would like answered. But despite what amounts to begging on the part of some analysts who cover Goldman, the firm won't divulge much. On the conference calls that follow the announcement of Goldman's quarterly results, CFO David Viniar refuses, again and again, to offer any guidance. "If any profitable opportunity to expand any of our businesses arose, we would take advantage of it, and if it didn't, we wouldn't," was a recent, relatively loquacious comment.

You can construe this as arrogance, as deliberate obfuscation--or more likely, as honesty. Paulson, likewise, won't hazard guesses about the firm's earnings. When Goldman was marketing its IPO, the CEO bluntly told investors that if they wanted stable earnings they should buy a toothpaste manufacturer. Around the same time Goldman did a strategic plan for its board--though it turned out to be way off in its projections. "Boy, were we wrong," Paulson laughs.

In his office on the 30th floor, Viniar, a Bronx native who joined Goldman in 1980 and worked as an investment banker until 1992, explains the exhaustive measures he takes to ensure that the firm values trades correctly. But when you ask how an outsider could verify that Goldman is marking its books correctly, he replies, "You can't." An investor recently asked how he could see how much risk the firm was taking. "I gave them the worst answer," Viniar says. "There's nothing you can look at." How does Viniar know it won't all go poof one day? "I think we have the right metrics and the right people," he says, "but I still can't look you in the eye and tell you nothing can go wrong. It can."

And yet it's not an accident that Goldman has kept churning out immense sums of money. One explanation is that the firm, for all its swashbuckling image, has begun to rake in bucks in a number of surprisingly mundane but immensely profitable ways. For instance, Goldman was late to the business of money management, and at the end of 1997 had less than one-third the assets under management that Merrill Lynch did and less than one-half what Morgan Stanley had. Today its partners Peter Kraus and Eric Schwartz run a team that manages $415 billion--a figure that rivals Merrill's $488 billion and Morgan's $500 billion. In addition, Goldman, along with Morgan Stanley, is a leader in what's known as "prime brokerage"--providing the exploding hedge fund industry with everything from back-office support to margin loans. This is not necessarily glamorous stuff, but it is a gold mine. These two businesses produced pretax profits of almost $1 billion in 2003, and that does not include the commissions that Goldman's hedge fund clients generate.

Another explanation is that Goldman's instincts have been on target, not just in taking advantage of low interest rates in fixed-income trading and exceptional volatility in commodities, but in all sorts of ways. "There are very few opportunities we've missed," says Paulson. In recent years Goldman has snapped up power plants in the U.S. and distressed debt in Asia. (The firm is now the largest owner of Japanese golf courses; at one point it owned one out of three cars and one out of seven pickup trucks in Thailand.) In 2002, Goldman invested $1.3 billion in Japan's Sumitomo Bank in a complicated deal that helped the firm extend loans to corporate clients. To date, the investment has added an average of 11% to the firm's operating earnings, says analyst Michael Hecht at Bank of America. (Of course, the investment could turn sour. Hecht estimates that Sumitomo, which sank in value over the summer, will lower Goldman's third-quarter pretax results by about $500 million.) Overall, what Goldman calls "trading and principal investing" has made $8.5 billion over the past 22 years, or 75% of Goldman's total pretax profits.

And that brings us to the third reason Goldman has made money, which is the way the firm thinks about risk. Some investors are wary of Goldman because a measure of the firm's risk taking, value-at-risk (which purports to measure the amount the firm could lose in one day of trading), has almost tripled since late 2000, to $69 million. Risk isn't an idle concern. Losses in FICC were the cause of one near-death experience for Goldman in 1994 and a coronary in 1998. After 1998 the bankers, who were then making all the money, at their annual winter conference put on a skit about FICC called "Fitanic."

No one at Goldman will deny that the firm takes risks. But it's hard to argue that Goldman is cavalier about it. Viniar, who as CFO is also head of risk management, sits two offices down from Blankfein, and Blankfein pops in--some 20 times a day. Goldman's risk management models collect, on average, 25,000 inputs a day to measure market risk. And Viniar is constantly running Goldman's positions through various scenarios. Among them: The "fall of 1998 stress test" (it mimics what happened that disastrous autumn), the "equities crash test" (this measures what would happen if the S&P 500 fell 50%), and the "Armageddon scenario" (self-explanatory). Because not having access to funds is the biggest risk for any trading firm, Goldman keeps what treasurer Beshel refers to as the "BONY box"--between $35 billion and $40 billion in short-term securities at the Bank of New York. "It's the No. 1 thing we do to help us sleep at night," she says.

Around Goldman, the question isn't whether the firm is taking too much risk but rather whether it's taking enough. "Our value-at-risk is up 40%, but our capital base is up 400%," says Blankfein. "Are we doing our jobs?" Adds Deighton: "The worst thing that could ever happen is for us to get gun-shy. The biggest risk we have is if we're sitting here with our overhead ticking away and nothing is happening." Ask Viniar if Goldman took enough risk over the past few years. He says: "You asked the easy question. I always wish we did more of what went well. The hard question is what it means for the future."

Goldman Sachs is a faith stock in some ways. And Wall Street's skeptics don't want to take anything on faith. That's partly why Goldman, like its peers, generally trades at a low multiple of both earnings and book value. The lack of faith is particularly strong today. While Goldman may be able to take advantage of the opportunities the market presents in an almost uncanny way, even it can't create opportunities where there aren't any--and lately the market has been tough.

But there are also specific concerns. Analysts want to know how much Goldman makes from so-called proprietary trading--buying and selling for its own account--vs. doing so on behalf of its clients. The problem is that it isn't always easy to distinguish between the two, even for people at Goldman. In markets such as fixed income and commodities, the firm often has to commit its own capital to facilitate a client trade--for instance, Goldman may buy bonds from a client that wants to sell. Once Goldman has its own capital in the game, there's a proprietary aspect to the trading, even though it may have been driven by a client's request. Goldman is trying to make money, after all. And from a risk point of view, it doesn't much matter why the firm has its capital at stake--just that it does. That said, Goldman does do some purely proprietary trading, as it has since even before the days of Bob Rubin, which the firm could detail. It won't.

Another worry is the future of Goldman's diminished investment-banking business. This is critical, because the market awards a higher multiple to those earnings than it does to trading earnings. Profits from trading are perceived as risky and unsustainable. Many Wall Streeters go so far as to call Goldman a hedge fund with a residual investment bank attached.

Blankfein argues that profits from Goldman's banking business understate its real value, because the relationships the firm has from its banking side provide its traders with opportunities that are unavailable to other firms--let alone hedge funds. "Our trading revenues are nowhere near as unstable as they would be at a hedge fund," he says. "Our traders come in every day and clients call them up with ideas." Whether you buy this point or not, it's indisputable that trading requires high levels of capital commitments, meaning that in flush times, banking earns a higher return on equity than trading ever could.

And it's unclear when, if ever, investment banking will again account for a substantial portion of Goldman's earnings. Banks like Citigroup, J.P. Morgan, and Bank of America have used their huge balance sheets to make inroads not just into debt underwriting, which pays lower fees, but also into higher-margin businesses such as equity underwriting and M&A. Citigroup's Chuck Prince has told analysts that his goal is to depose Goldman from its top M&A ranking over the next three years. Other bank CEOs are equally confident that their ability to combine lending and traditional investment-banking products is the model that will win. "I view Goldman Sachs as the Charge of the Light Brigade," says one rival executive. "Their horse is prancing up the mountain looking for the high-value-added stuff, while we're taking the lower stuff."

Goldman bankers know they're under siege. They talk about how vicious the competition is, how hard it is to be a senior banker today. "The last three years have been more difficult than any I can remember--ever," says Levy, who is the co-head of Goldman's M&A department. "This is an environment where we cannot afford to make a mistake."

But it's difficult to tell if this is Goldman paranoia or something else, because behind the doomsday words are the numbers, and the numbers don't tell such a scary story. Yes, Goldman has lost market share in debt underwriting, but the firm's market share in equity and M&A, while down from the boom years, is higher than it was in the mid-1990s--and Goldman is still No. 1 (or a close No. 2) in every category, according to Thomson Financial. "Three years ago everyone was saying Goldman Sachs was dead because the balance sheet was everything," says Tom Finucane, an analyst at State Street Research, which owns Goldman stock. "Look at who's still at the top of the league tables." But even if Goldman can keep its lead, it's hard to imagine that the bonanza days of the last bull market will return anytime soon.

You might think, given all the scrutiny and the skepticism, that there would be some regret about the IPO, some yearning for the old days when the firm could slam the door on the outside world. But if there is, it's hidden well, even by those who say they were opposed to going public. The memory of 1994, when partners fled in droves, yanking their capital with them and greatly exacerbating the pressure that the trading losses caused, is still fresh. Current partners like Suzanne Nora Johnson, the head of the firm's research department, argue that there's no way Goldman's current 253 partners could have shouldered the weight of the last downturn. Cohn offers another explanation: "Had we not gone public, we would be a boutique. And none of us want to be in a boutique."

But as a large public company, Goldman will face bigger questions than its stock price in the next few years. Can Goldman continue to be Goldman? Can the firm succeed in indoctrinating thousands upon thousands of people into that Goldman culture without drowning under the weight of formal processes and bureaucracy? As of last June, Goldman insiders were free to sell the last of the shares that were locked up after the IPO; insider ownership has already declined from more than 80% after the IPO to just about 30% today. Admits one partner: "The culture is as stressed as it's ever been."

Talk to others on Wall Street, and you'll hear that there's a "brain drain" at Goldman, that the turnover isn't just Goldman's ruthlessness at work but rather people getting out while the getting is good. The days when Goldman partners made more than anyone on Wall Street are long gone. "The ring of being a partner isn't as golden as it used to be," says one former partner. And among those who have left, the mood is often dark. Goldman has become too bureaucratic and too political, they say. You can have more fun and make more money elsewhere. As one person who joined the firm upon graduating from Princeton and spent more than a decade there put it, "Goldman has gone from being the last place you'll ever work to just another stop in your career."

While the firm has a number of programs to promote diversity, some women who have left say that Goldman has missed opportunities to keep people who could have continued to contribute greatly. "They make investments in women and then just write them off," says a former senior employee. Of the 14 female partners at the time of the IPO, in fact, only four remain. (The author of this story was an analyst in Goldman's investment-banking division from 1992 to 1995, before she left to become a reporter. She would most certainly not have made partner--but that's not Goldman's fault.)

In all likelihood, it won't be obvious for years whether Goldman has lost something special. Most of its senior people came up through the partnership, and cultures don't change overnight. But in a way the answer is clear already: Goldman can't be what it was. "Every year the firm becomes bigger, it looks less like the firm it did," says a former partner. "It doesn't mean Goldman won't be successful. There are incredible opportunities. But every year it's less of a partnership."

Another question is whether Goldman really deserves its pristine reputation--or whether, as some argue, ethical stains simply don't stick to its Teflon surface. The evidence is mixed. There is substance behind all the talk about the importance of reputation that you'll hear around Goldman (and you will hear a lot of that). Goldman was a major shipper of natural gas into California and a major player in electricity. Yet amid all the aftershocks emanating from California's 2000--01 power crisis, there haven't been headlines about the firm--mainly because Goldman curtailed its business ahead of the crisis. "We knew we didn't want to have anything to do with it," Paulson says. "When you're trading in a system that is so flawed, you may make a lot of money, but you know a lot of mud is going to splash."

There are other facts that make Goldman's talk sound like so many words. In late July, Goldman, along with three other Wall Street firms, agreed to pay $5 million to settle charges that investors paid huge markups on bond purchases or received too little when they sold. (The investigation of abuses in the bond market is ongoing.) Former Goldman economist John Youngdahl pled guilty to federal securities fraud for passing embargoed information to Goldman traders when the government halted the sale of the 30-year bond; Goldman itself paid $9.3 million for failing to prevent "the misuse of material nonpublic information." And Spear Leeds, the NYSE specialist firm that Goldman bought in 2000, agreed in 2004 to pay $45.3 million to settle charges that it traded ahead of customers.

Goldman's biggest opportunity to stand behind its talk was during the boom years. And it missed the chance. It paid $110 million in 2003 for its share of the Wall Street global settlement over research improprieties. (Goldman, unlike some of its peers, wasn't accused of fraudulent research.) The firm also collected hundreds of millions in fees for underwriting a slew of companies that soon went bankrupt, such as PlanetRx.com, Webvan, and more.

Goldman executives get quite agitated about the firm's behavior during the go-go years. "Saying no is one of the most powerful things we can do," notes investment-banking co-head Scott Kapnick, whose father, Harvey, was the CEO of Arthur Andersen in the 1970s. "The lessons of the last three years are not lost on anyone here," says Kennedy, who is the co-chair of the commitments committee, which decides what business Goldman will do. "It was the low point in my career. I am personally embarrassed."

To his credit, Paulson has put in place a compulsory program for the entire firm on compliance and reputational judgment. And he has told people that the slightest infraction will cost them their jobs. But as the CEO admits, it's impossible to stop every indiscretion: "We will never eliminate people doing bad things. In a town of 20,000 people, there's a jail."

Indeed, Goldman's reputation is the heart of its future. But as Paulson well knows, reputation isn't just a matter of staying within the letter of the law--or managing to escape the next Eliot Spitzer investigation. Goldman must also navigate an ever more complex, more competitive world, one where very few rules are as straightforward as they seem--and where the currency is something as nuanced as information.

Goldman's first principle is that its "clients' interests always come first," but on Wall Street that's not a simple precept by which to live. "This is a business of managing conflicts," says Blankfein. The interests of Goldman's clients often conflict with each other--and with Goldman's. For instance, a private equity firm may be a client of Goldman's investment bank or a competitor to Goldman's private equity operation. In a simple example, if Goldman finds a great investment, does Goldman put its own money in the deal, or does the firm tell a client? This inherent competition is accepted by some--"We're all big boys," says Coulter--but not by others. "It drives me nuts," says one client. "I scream at them once a month."

Goldman's massive trading business presents a more challenging conundrum: Where is the line that separates the firm's use of the information it gets about its clients' trading activities and the abuse of that information? As a brokerage firm, Goldman can have more insight into the market than its clients do. As analyst Brad Hintz at Sanford Bernstein explains, "It's like you're walking along and you kick an anthill, and you see all the ants scurrying right and left. You can see it, but the ants can't. That's the difference between being a brokerage firm and being a client." There's nothing illegal about this: It's the way the Street has always worked.

What's new is the amount of capital that Goldman and its peers have in the game, either through purely proprietary trading or through facilitating their clients' trades. Even equities trading, which used to be a gentlemanly business where brokers collected commissions but didn't risk their own capital, is changing. As commissions plummet, analysts such as Hintz argue that over time, Goldman and other brokerages will make their money by trading more and more for their own accounts.

Some privately complain that Goldman is aggressive about using information to make profits for itself, especially in less transparent markets like commodities and distressed debt. How much of this is real and how much is perception, paranoia, and jealousy? It's impossible to tell. Goldman says it has Chinese walls around prime brokerage and proprietary trading to prevent the misuse of information. Huge customers often have market clout and insight, too, and they don't hesitate to beat Goldman over the head with it. And even the people who complain the most vociferously usually concede that they keep doing business with the firm.

This is another contradiction at the heart of the modern Goldman. Its omnipresence in the markets inspires resentment and makes it prone to conflicts. But this reach also can make it difficult for people to refuse to do business with it. That's the thing about being the money machine: Everyone needs you. And, of course, it's a virtuous circle that explains Goldman's success: The bigger and more powerful the firm gets, the more clients need it.

If something happens to destroy Goldman's fortunes, chances are it will be something that no one saw coming. It will be an oh-so-slow erosion of that virtuous circle--or something sudden and dramatic: a business practice, for instance, that seems acceptable today but doesn't tomorrow, or an event that shakes the country or the world.

"Look, fortunes are going to be made," says Blankfein. "But the biggest risk is what you can't see today. That's why life is the ability to step outside yourself, to see that something that you can't see today, something that one day you'll hit yourself on the forehead and say, 'Why didn't I see it that way?' You won't be able to imagine not having seen it, except that you'll remember not having seen it."

The overarching uncertainty is enough to make even the pragmatic Paulson sound, well, almost philosophical: "The one thing I know for sure is that we are related to everything in the world. If there's a problem that impacts the world negatively, it will impact us negatively. The only question is, How much?"

FEEDBACK bmclean@fortunemail.com